Compare commits

...

34 Commits

Author SHA1 Message Date
Glenn Randers-Pehrson
448092b8eb [libpng17] Imported from libpng-1.7.0beta24.tar 2013-12-15 11:22:49 -06:00
Glenn Randers-Pehrson
1faa6ff32c [libpng17] 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().
2013-12-15 08:47:52 -06:00
John Bowler
3adf438f39 [libpng17] 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.
2013-12-13 23:55:41 -06:00
John Bowler
d45a632529 [libpng17] Bookkeeping: Moved functions around (no changes). Moved transform
function definitions before the place where they are called so that
they can be masde 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
2013-12-13 23:50:30 -06:00
John Bowler
5a1b8d106d [libpng17] Tidied up pngfix inits and fixed pngtest no-write builds. 2013-12-01 15:50:03 -06:00
Glenn Randers-Pehrson
05f896004f [libpng17] 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
2013-11-28 16:14:03 -06:00
Glenn Randers-Pehrson
a4f1de9109 [libpng17] Avoid a scan-build warning in pngtest.c 2013-11-25 13:13:32 -06:00
Glenn Randers-Pehrson
7ebb353944 [libpng17] Removed or marked PNG_UNUSED some harmless "dead assignments"
reported by clang scan-build.
2013-11-25 11:08:56 -06:00
Glenn Randers-Pehrson
b1f0c7d0e5 [libpng17] Revised manual about handling of zlib CMF 2013-11-24 13:31:26 -06:00
Glenn Randers-Pehrson
55df049962 [libpng17] Updated CHANGES and ANNOUNCE 2013-11-23 19:26:57 -06:00
Glenn Randers-Pehrson
219ba1370e [libpng17] Bump version to 1.7.0beta24 2013-11-23 19:02:56 -06:00
Glenn Randers-Pehrson
4dc4471bb5 [libpng17] Imported from libpng-1.7.0beta23.tar 2013-11-23 19:01:01 -06:00
Glenn Randers-Pehrson
a199a15b28 [libpng17] Added another "assert()" to pngfix.c 2013-11-23 12:58:32 -06:00
Glenn Randers-Pehrson
e9446f2f95 [libpng17] Added an "assert()" to pngfix.c to quiet a clang scanner warning. 2013-11-23 12:14:02 -06:00
John Bowler
f70f2fcf48 [libpng17] Finished merging with libpng16 (pngrutil.c and pnglibconf.dfa) 2013-11-23 08:20:39 -06:00
John Bowler
158c947ce2 [libpng17] Revert recent change to pngvalid.c. 2013-11-22 18:34:37 -06:00
John Bowler
7e2707e2af [libpng17] Ported recent changes from libpng16 to libpng17
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.
2013-11-22 18:01:30 -06:00
Glenn Randers-Pehrson
81824079b5 [libpng17] Added "-Wall" to CFLAGS in contrib/pngminim/*/makefile 2013-11-18 17:03:24 -06:00
Glenn Randers-Pehrson
e8e1add0a8 [libpng17] Moved prototype for png_handle_unknown() in pngpriv.h outside of
the #ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED/#endif block.
2013-11-18 15:35:55 -06:00
Glenn Randers-Pehrson
a549223f91 [libpng17] Added #ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED/#endif in pngpread.c 2013-11-17 11:09:57 -06:00
Glenn Randers-Pehrson
ebe7945b96 [libpng17] Bump version to 1.7.0beta23 2013-11-07 08:33:41 -06:00
Glenn Randers-Pehrson
e3e35aa980 [libpng17] Imported from libpng-1.7.0beta22.tar 2013-11-07 08:33:31 -06:00
John Bowler
9f2fbcf5d8 [libpng17] Fixed #include in filter_neon_intrinsics.c and ctype macros. The
ctype char checking macros take an unsigned char argument, not a signed char.
2013-11-04 13:50:20 -06:00
Glenn Randers-Pehrson
264ba68afe [libpng17] Restored a line to the libpng-1.7.0beta21 entry in ANNOUNCE 2013-11-03 20:56:03 -06:00
Glenn Randers-Pehrson
3ef99c9316 [libpng17] Bump version to 1.7.0beta22 2013-11-02 15:39:27 -05:00
John Bowler
4b29f97f9b [libpng17] Port recent libpng-1.6.7beta03, beta04 changes to 1.7.0beta21
(mainly ARMv8 support)
2013-11-02 15:29:45 -05:00
Glenn Randers-Pehrson
fe6e6cf130 [libpng17] Fixed some typos in autogen.s that don't affect its behavior 2013-10-13 17:12:44 -05:00
Glenn Randers-Pehrson
547df4b9ca [libpng17] Bump version to 1.7.0beta21 2013-10-13 14:10:23 -05:00
Glenn Randers-Pehrson
a0d4a0b971 [libpng17] Imported from libpng-1.7.0beta20.tar 2013-10-13 14:02:50 -05:00
John Bowler
72c2577302 [libpng17] 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.
2013-10-06 09:02:09 -05:00
John Bowler
091fab713c [libpng17] 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).
2013-10-06 08:51:42 -05:00
John Bowler
32aeb2cd4e [libpng17] 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
2013-10-06 08:47:43 -05:00
John Bowler
c1bcdc3a86 [libpng17] 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.
2013-10-06 08:18:31 -05:00
Glenn Randers-Pehrson
553b8f85d3 [libpng17] Bump version to 1.7.0beta20 2013-10-02 13:47:20 -05:00
57 changed files with 2124 additions and 1072 deletions

103
ANNOUNCE
View File

@@ -1,5 +1,5 @@
Libpng 1.7.0beta19 - September 30, 2013
Libpng 1.7.0beta24 - December 15, 2013
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.
@@ -9,19 +9,19 @@ Files available for download:
Source files with LF line endings (for Unix/Linux) and with a
"configure" script
1.7.0beta19.tar.xz (LZMA-compressed, recommended)
1.7.0beta19.tar.gz
1.7.0beta24.tar.xz (LZMA-compressed, recommended)
1.7.0beta24.tar.gz
Source files with CRLF line endings (for Windows), without the
"configure" script
lp170b19.7z (LZMA-compressed, recommended)
lp170b19.zip
lp170b24.7z (LZMA-compressed, recommended)
lp170b24.zip
Other information:
1.7.0beta19-README.txt
1.7.0beta19-LICENSE.txt
1.7.0beta24-README.txt
1.7.0beta24-LICENSE.txt
Gnupg/*.asc (PGP armored detached signatures)
Changes since the last public release (1.6.0):
@@ -372,6 +372,95 @@ Version 1.7.0beta19 [September 30, 2013]
png_modifier, but this change avoids the warning.
Fixed some grammatical changes and updated the png chunks list.
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 1.7.0beta21 [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 1.7.0beta22 [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 1.7.0beta23 [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 1.7.0beta24 [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 masde 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().
Send comments/corrections/commendations to png-mng-implement at lists.sf.net
(subscription required; visit
https://lists.sourceforge.net/lists/listinfo/png-mng-implement

89
CHANGES
View File

@@ -4661,6 +4661,95 @@ Version 1.7.0beta19 [September 30, 2013]
png_modifier, but this change avoids the warning.
Fixed some grammatical changes and updated the png chunks list.
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 1.7.0beta21 [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 1.7.0beta22 [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 1.7.0beta23 [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 1.7.0beta24 [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 masde 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().
Send comments/corrections/commendations to png-mng-implement at lists.sf.net
(subscription required; visit
https://lists.sourceforge.net/lists/listinfo/png-mng-implement

View File

@@ -252,7 +252,7 @@ endif(NOT WIN32 OR CYGWIN OR MINGW)
# SET UP LINKS
if(PNG_SHARED)
set_target_properties(${PNG_LIB_NAME} PROPERTIES
# VERSION 17.${PNGLIB_RELEASE}.1.7.0beta19
# VERSION 17.${PNGLIB_RELEASE}.1.7.0beta24
VERSION 17.${PNGLIB_RELEASE}.0
SOVERSION 17
CLEAN_DIRECT_OUTPUT 1)

View File

@@ -10,7 +10,7 @@ this sentence.
This code is released under the libpng license.
libpng versions 1.2.6, August 15, 2004, through 1.7.0beta19, September 30, 2013, are
libpng versions 1.2.6, August 15, 2004, through 1.7.0beta24, December 15, 2013, are
Copyright (c) 2004, 2006-2013 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 30, 2013
December 15, 2013

View File

@@ -12,6 +12,16 @@ check_PROGRAMS= pngtest pngunknown pngstest pngvalid
# 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
@@ -72,7 +82,7 @@ libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_SOURCES = png.c pngerror.c\
if PNG_ARM_NEON
libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_SOURCES += arm/arm_init.c\
arm/filter_neon.S
arm/filter_neon.S arm/filter_neon_intrinsics.c
endif
nodist_libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_SOURCES = pnglibconf.h
@@ -196,9 +206,21 @@ $(srcdir)/scripts/pnglibconf.h.prebuilt:
# The following is necessary to ensure that the local pnglibconf.h is used, not
# an installed one (this can happen immediately after on a clean system if
# 'make test' is the first thing the user does.)
pngstest.o pngvalid.o pngtest.o pngunknown.o timepng.o: pnglibconf.h
pngfix.o png-fix-itxt.o: pnglibconf.h
# 'make test' is the first thing the user does.) Only files which include
# one of the png source files (typically png.h or pngpriv.h) need to be listed
# here:
pngtest.o: pnglibconf.h
contrib/libtests/makepng.o: pnglibconf.h
contrib/libtests/pngstest.o: pnglibconf.h
contrib/libtests/pngunknown.o: pnglibconf.h
contrib/libtests/pngvalid.o: pnglibconf.h
contrib/libtests/readpng.o: pnglibconf.h
contrib/libtests/tarith.o: pnglibconf.h
contrib/libtests/timepng.o: pnglibconf.h
contrib/tools/makesRGB.o: pnglibconf.h
contrib/tools/pngfix.o: pnglibconf.h
# We must use -DPNG_NO_USE_READ_MACROS here even when the library may actually
# be built with PNG_USE_READ_MACROS; this prevents the read macros from
@@ -304,3 +326,8 @@ uninstall-hook:
rm -f '$(DESTDIR)$(libdir)/libpng.sl'
rm -f '$(DESTDIR)$(libdir)/libpng.dylib'
rm -f '$(DESTDIR)$(libdir)/libpng.dll.a'
# 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)

2
README
View File

@@ -1,4 +1,4 @@
README for libpng version 1.7.0beta19 - September 30, 2013 (shared library 17.0)
README for libpng version 1.7.0beta24 - December 15, 2013 (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.

View File

@@ -3,7 +3,7 @@
*
* Copyright (c) 2013 Glenn Randers-Pehrson
* Written by Mans Rullgard, 2011.
* Last changed in libpng 1.5.17 [July 18, 2013]
* Last changed in libpng 1.6.8 [(PENDING RELEASE)]
*
* This code is released under the libpng license.
* For conditions of distribution and use, see the disclaimer
@@ -11,7 +11,8 @@
*/
/* This is required to get the symbol renames, which are #defines, and also
* includes the definition (or not) of PNG_ARM_NEON_OPT.
* includes the definition (or not) of PNG_ARM_NEON_OPT and
* PNG_ARM_NEON_IMPLEMENTATION.
*/
#define PNG_VERSION_INFO_ONLY
#include "../pngpriv.h"
@@ -20,6 +21,13 @@
.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
@@ -235,3 +243,4 @@ func png_read_filter_row_paeth3_neon, export=1
endfunc
#endif /* PNG_ARM_NEON_OPT > 0 */
#endif /* PNG_READ_SUPPORTED */
#endif /* PNG_ARM_NEON_IMPLEMENTATION == 2 (assembler) */

View File

@@ -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 [(PENDING RELEASE)]
*
* 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) */

View File

@@ -9,15 +9,15 @@
#
# 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 rune autogen.sh because ./configure --enable-maintainer-mode says
# what to do if Makeile.am or configure.ac are changed.
# 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 for (a) use while testing the script and (b) to fix up systems that
# 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.
@@ -76,6 +76,10 @@ done
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"
@@ -94,6 +98,7 @@ libpng_configure_dirs=".deps"
# 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 $libpng_autotools_extra
}
clean_configure(){
@@ -198,7 +203,7 @@ case "$mode" in
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 " compatiblity with prior releases."
echo " compatibility with prior releases."
exit 1
else
exec >&2

View File

@@ -18,7 +18,7 @@ AC_PREREQ([2.68])
dnl Version number stuff here:
AC_INIT([libpng],[1.7.0beta19],[png-mng-implement@lists.sourceforge.net])
AC_INIT([libpng],[1.7.0beta24],[png-mng-implement@lists.sourceforge.net])
AC_CONFIG_MACRO_DIR([scripts])
# libpng does not follow GNU file name conventions (hence 'foreign')
@@ -27,7 +27,7 @@ AC_CONFIG_MACRO_DIR([scripts])
# 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])
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
@@ -39,7 +39,7 @@ 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=1.7.0beta19
PNGLIB_VERSION=1.7.0beta24
PNGLIB_MAJOR=1
PNGLIB_MINOR=7
PNGLIB_RELEASE=0

46
contrib/conftest/README Normal file
View File

@@ -0,0 +1,46 @@
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.
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.

58
contrib/conftest/read.dfa Normal file
View File

@@ -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

View File

@@ -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

View File

@@ -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

View File

@@ -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

View File

@@ -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')

View File

@@ -517,14 +517,23 @@ static void format_default(format_list *pf, int redundant)
for (f=0; f<FORMAT_COUNT; ++f)
{
/* Eliminate redundant settings. */
/* BGR is meaningless if no color: */
if ((f & PNG_FORMAT_FLAG_COLOR) == 0 && (f & PNG_FORMAT_FLAG_BGR) != 0)
/* Eliminate redundant and unsupported settings. */
# ifdef PNG_FORMAT_BGR_SUPPORTED
/* BGR is meaningless if no color: */
if ((f & PNG_FORMAT_FLAG_COLOR) == 0 &&
(f & PNG_FORMAT_FLAG_BGR) != 0)
# else
if ((f & 0x10U/*HACK: fixed value*/) != 0)
# endif
continue;
/* AFIRST is meaningless if no alpha: */
if ((f & PNG_FORMAT_FLAG_ALPHA) == 0 &&
(f & PNG_FORMAT_FLAG_AFIRST) != 0)
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if ((f & PNG_FORMAT_FLAG_ALPHA) == 0 &&
(f & PNG_FORMAT_FLAG_AFIRST) != 0)
# else
if ((f & 0x20U/*HACK: fixed value*/) != 0)
# endif
continue;
format_set(pf, f);
@@ -786,6 +795,7 @@ gp_ga8(Pixel *p, png_const_voidp pb)
p->a = pp[1];
}
#ifdef PNG_FORMAT_AFIRST_SUPPORTED
static void
gp_ag8(Pixel *p, png_const_voidp pb)
{
@@ -794,6 +804,7 @@ gp_ag8(Pixel *p, png_const_voidp pb)
p->r = p->g = p->b = pp[1];
p->a = pp[0];
}
#endif
static void
gp_rgb8(Pixel *p, png_const_voidp pb)
@@ -806,6 +817,7 @@ gp_rgb8(Pixel *p, png_const_voidp pb)
p->a = 255;
}
#ifdef PNG_FORMAT_BGR_SUPPORTED
static void
gp_bgr8(Pixel *p, png_const_voidp pb)
{
@@ -816,6 +828,7 @@ gp_bgr8(Pixel *p, png_const_voidp pb)
p->b = pp[0];
p->a = 255;
}
#endif
static void
gp_rgba8(Pixel *p, png_const_voidp pb)
@@ -828,6 +841,7 @@ gp_rgba8(Pixel *p, png_const_voidp pb)
p->a = pp[3];
}
#ifdef PNG_FORMAT_BGR_SUPPORTED
static void
gp_bgra8(Pixel *p, png_const_voidp pb)
{
@@ -838,7 +852,9 @@ gp_bgra8(Pixel *p, png_const_voidp pb)
p->b = pp[0];
p->a = pp[3];
}
#endif
#ifdef PNG_FORMAT_AFIRST_SUPPORTED
static void
gp_argb8(Pixel *p, png_const_voidp pb)
{
@@ -849,7 +865,9 @@ gp_argb8(Pixel *p, png_const_voidp pb)
p->b = pp[3];
p->a = pp[0];
}
#endif
#if defined(PNG_FORMAT_AFIRST_SUPPORTED) && defined(PNG_FORMAT_BGR_SUPPORTED)
static void
gp_abgr8(Pixel *p, png_const_voidp pb)
{
@@ -860,6 +878,7 @@ gp_abgr8(Pixel *p, png_const_voidp pb)
p->b = pp[1];
p->a = pp[0];
}
#endif
static void
gp_g16(Pixel *p, png_const_voidp pb)
@@ -879,6 +898,7 @@ gp_ga16(Pixel *p, png_const_voidp pb)
p->a = pp[1];
}
#ifdef PNG_FORMAT_AFIRST_SUPPORTED
static void
gp_ag16(Pixel *p, png_const_voidp pb)
{
@@ -887,6 +907,7 @@ gp_ag16(Pixel *p, png_const_voidp pb)
p->r = p->g = p->b = pp[1];
p->a = pp[0];
}
#endif
static void
gp_rgb16(Pixel *p, png_const_voidp pb)
@@ -899,6 +920,7 @@ gp_rgb16(Pixel *p, png_const_voidp pb)
p->a = 65535;
}
#ifdef PNG_FORMAT_BGR_SUPPORTED
static void
gp_bgr16(Pixel *p, png_const_voidp pb)
{
@@ -909,6 +931,7 @@ gp_bgr16(Pixel *p, png_const_voidp pb)
p->b = pp[0];
p->a = 65535;
}
#endif
static void
gp_rgba16(Pixel *p, png_const_voidp pb)
@@ -921,6 +944,7 @@ gp_rgba16(Pixel *p, png_const_voidp pb)
p->a = pp[3];
}
#ifdef PNG_FORMAT_BGR_SUPPORTED
static void
gp_bgra16(Pixel *p, png_const_voidp pb)
{
@@ -931,7 +955,9 @@ gp_bgra16(Pixel *p, png_const_voidp pb)
p->b = pp[0];
p->a = pp[3];
}
#endif
#ifdef PNG_FORMAT_AFIRST_SUPPORTED
static void
gp_argb16(Pixel *p, png_const_voidp pb)
{
@@ -942,7 +968,9 @@ gp_argb16(Pixel *p, png_const_voidp pb)
p->b = pp[3];
p->a = pp[0];
}
#endif
#if defined(PNG_FORMAT_AFIRST_SUPPORTED) && defined(PNG_FORMAT_BGR_SUPPORTED)
static void
gp_abgr16(Pixel *p, png_const_voidp pb)
{
@@ -953,6 +981,7 @@ gp_abgr16(Pixel *p, png_const_voidp pb)
p->b = pp[1];
p->a = pp[0];
}
#endif
/* Given a format, return the correct one of the above functions. */
static void (*
@@ -966,29 +995,35 @@ get_pixel(png_uint_32 format))(Pixel *p, png_const_voidp pb)
{
if (format & PNG_FORMAT_FLAG_COLOR)
{
if (format & PNG_FORMAT_FLAG_BGR)
{
if (format & PNG_FORMAT_FLAG_ALPHA)
# ifdef PNG_FORMAT_BGR_SUPPORTED
if (format & PNG_FORMAT_FLAG_BGR)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_abgr16;
if (format & PNG_FORMAT_FLAG_ALPHA)
{
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_abgr16;
else
# endif
return gp_bgra16;
}
else
return gp_bgra16;
return gp_bgr16;
}
else
return gp_bgr16;
}
else
# endif
{
if (format & PNG_FORMAT_FLAG_ALPHA)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_argb16;
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_argb16;
else
else
# endif
return gp_rgba16;
}
@@ -1001,10 +1036,12 @@ get_pixel(png_uint_32 format))(Pixel *p, png_const_voidp pb)
{
if (format & PNG_FORMAT_FLAG_ALPHA)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_ag16;
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_ag16;
else
else
# endif
return gp_ga16;
}
@@ -1017,29 +1054,35 @@ get_pixel(png_uint_32 format))(Pixel *p, png_const_voidp pb)
{
if (format & PNG_FORMAT_FLAG_COLOR)
{
if (format & PNG_FORMAT_FLAG_BGR)
{
if (format & PNG_FORMAT_FLAG_ALPHA)
# ifdef PNG_FORMAT_BGR_SUPPORTED
if (format & PNG_FORMAT_FLAG_BGR)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_abgr8;
if (format & PNG_FORMAT_FLAG_ALPHA)
{
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_abgr8;
else
# endif
return gp_bgra8;
}
else
return gp_bgra8;
return gp_bgr8;
}
else
return gp_bgr8;
}
else
# endif
{
if (format & PNG_FORMAT_FLAG_ALPHA)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_argb8;
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_argb8;
else
else
# endif
return gp_rgba8;
}
@@ -1052,10 +1095,12 @@ get_pixel(png_uint_32 format))(Pixel *p, png_const_voidp pb)
{
if (format & PNG_FORMAT_FLAG_ALPHA)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_ag8;
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
return gp_ag8;
else
else
# endif
return gp_ga8;
}
@@ -2618,13 +2663,15 @@ component_loc(png_byte loc[4], png_uint_32 format)
loc[2] = 1;
if (format & PNG_FORMAT_FLAG_BGR)
{
loc[1] = 2;
loc[3] = 0;
}
# ifdef PNG_FORMAT_BGR_SUPPORTED
if (format & PNG_FORMAT_FLAG_BGR)
{
loc[1] = 2;
loc[3] = 0;
}
else
else
# endif
{
loc[1] = 0;
loc[3] = 2;
@@ -2639,15 +2686,17 @@ component_loc(png_byte loc[4], png_uint_32 format)
if (format & PNG_FORMAT_FLAG_ALPHA)
{
if (format & PNG_FORMAT_FLAG_AFIRST)
{
loc[0] = 0;
++loc[1];
++loc[2];
++loc[3];
}
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
if (format & PNG_FORMAT_FLAG_AFIRST)
{
loc[0] = 0;
++loc[1];
++loc[2];
++loc[3];
}
else
else
# endif
loc[0] = channels;
++channels;
@@ -3017,17 +3066,25 @@ read_file(Image *image, png_uint_32 format, png_const_colorp background)
return logerror(image, "memory init: ", image->file_name, "");
}
else if (image->input_file != NULL)
{
if (!png_image_begin_read_from_stdio(&image->image, image->input_file))
return logerror(image, "stdio init: ", image->file_name, "");
}
# ifdef PNG_STDIO_SUPPORTED
else if (image->input_file != NULL)
{
if (!png_image_begin_read_from_stdio(&image->image, image->input_file))
return logerror(image, "stdio init: ", image->file_name, "");
}
else
{
if (!png_image_begin_read_from_file(&image->image, image->file_name))
return logerror(image, "file init: ", image->file_name, "");
}
else
{
if (!png_image_begin_read_from_file(&image->image, image->file_name))
return logerror(image, "file init: ", image->file_name, "");
}
# else
else
{
return logerror(image, "unsupported file/stdio init: ",
image->file_name, "");
}
# endif
/* This must be set after the begin_read call: */
if (image->opts & sRGB_16BIT)
@@ -3466,11 +3523,19 @@ main(int argc, char **argv)
memset(gpc_error_via_linear, 0, sizeof gpc_error_via_linear);
}
else if (strcmp(arg, "--file") == 0)
opts |= READ_FILE;
# ifdef PNG_STDIO_SUPPORTED
opts |= READ_FILE;
# else
return 77; /* skipped: no support */
# endif
else if (strcmp(arg, "--memory") == 0)
opts &= ~READ_FILE;
else if (strcmp(arg, "--stdio") == 0)
opts |= USE_STDIO;
# ifdef PNG_STDIO_SUPPORTED
opts |= USE_STDIO;
# else
return 77; /* skipped: no support */
# endif
else if (strcmp(arg, "--name") == 0)
opts &= ~USE_STDIO;
else if (strcmp(arg, "--verbose") == 0)

View File

@@ -866,13 +866,6 @@ check_handling(display *d, int def, png_uint_32 chunks, png_uint_32 known,
const char *type;
const char *errorx = NULL;
# ifdef PNG_READ_USER_CHUNKS_SUPPORTED
const char *callback = (set_callback ? ",callback" : "");
# else
# define callback ""
UNUSED(set_callback)
# endif
if (chunk_info[i].unknown)
{
if (keep == PNG_HANDLE_CHUNK_AS_DEFAULT)
@@ -970,7 +963,8 @@ check_handling(display *d, int def, png_uint_32 chunks, png_uint_32 known,
if (errorx != NULL)
{
++(d->error_count);
fprintf(stderr, "%s(%s%s): %s %s %s: %s\n", d->file, d->test, callback,
fprintf(stderr, "%s(%s%s): %s %s %s: %s\n", d->file, d->test,
set_callback ? ",callback" : "",
type, chunk_info[i].name, position, errorx);
}

View File

@@ -45,7 +45,10 @@
# include PNG_ZLIB_HEADER
#endif
#ifdef PNG_WRITE_SUPPORTED /* else pngvalid can do nothing */
/* pngvalid requires write support and one of the fixed or floating point APIs.
*/
#if defined(PNG_WRITE_SUPPORTED) &&\
(defined(PNG_FIXED_POINT_SUPPORTED) || defined(PNG_FLOATING_POINT_SUPPORTED))
#if PNG_LIBPNG_VER < 10500
/* This deliberately lacks the PNG_CONST. */
@@ -106,6 +109,11 @@ typedef png_byte *png_const_bytep;
# endif
#endif
/* Fixups for various minimal builds */
#ifndef PNG_ERROR_TEXT_SUPPORTED
# define png_error(a,b) png_err(a)
#endif
/***************************** EXCEPTION HANDLING *****************************/
#ifdef PNG_FREESTANDING_TESTS
# include <cexcept.h>
@@ -1268,6 +1276,7 @@ store_current_palette(png_store *ps, int *npalette)
#endif /* PNG_READ_SUPPORTED */
/***************************** MEMORY MANAGEMENT*** ***************************/
#ifdef PNG_USER_MEM_SUPPORTED
/* A store_memory is simply the header for an allocated block of memory. The
* pointer returned to libpng is just after the end of the header block, the
* allocated memory is followed by a second copy of the 'mark'.
@@ -1468,6 +1477,7 @@ store_free(png_structp ppIn, png_voidp memory)
this->next = NULL;
store_memory_free(pp, pool, this);
}
#endif /* PNG_USER_MEM_SUPPORTED */
/* Setup functions. */
/* Cleanup when aborting a write or after storing the new file. */
@@ -1493,7 +1503,9 @@ store_write_reset(png_store *ps)
/* And make sure that all the memory has been freed - this will output
* spurious errors in the case of memory corruption above, but this is safe.
*/
store_pool_delete(ps, &ps->write_memory_pool);
# ifdef PNG_USER_MEM_SUPPORTED
store_pool_delete(ps, &ps->write_memory_pool);
# endif
store_freenew(ps);
}
@@ -1517,16 +1529,20 @@ set_store_for_write(png_store *ps, png_infopp ppi,
store_write_reset(ps);
safecat(ps->wname, sizeof ps->wname, 0, name);
/* Don't do the slow memory checks if doing a speed test. */
if (ps->speed)
/* Don't do the slow memory checks if doing a speed test, also if user
* memory is not supported we can't do it anyway.
*/
# ifdef PNG_USER_MEM_SUPPORTED
if (!ps->speed)
ps->pwrite = png_create_write_struct_2(PNG_LIBPNG_VER_STRING,
ps, store_error, store_warning, &ps->write_memory_pool,
store_malloc, store_free);
else
# endif
ps->pwrite = png_create_write_struct(PNG_LIBPNG_VER_STRING,
ps, store_error, store_warning);
else
ps->pwrite = png_create_write_struct_2(PNG_LIBPNG_VER_STRING,
ps, store_error, store_warning, &ps->write_memory_pool,
store_malloc, store_free);
png_set_write_fn(ps->pwrite, ps, store_write, store_flush);
# ifdef PNG_SET_OPTION_SUPPORTED
@@ -1574,8 +1590,10 @@ store_read_reset(png_store *ps)
}
# endif
/* Always do this to be safe. */
store_pool_delete(ps, &ps->read_memory_pool);
# ifdef PNG_USER_MEM_SUPPORTED
/* Always do this to be safe. */
store_pool_delete(ps, &ps->read_memory_pool);
# endif
ps->current = NULL;
ps->next = NULL;
@@ -1635,14 +1653,16 @@ set_store_for_read(png_store *ps, png_infopp ppi, png_uint_32 id,
* However, given that store_error works correctly in these circumstances
* we don't ever expect NULL in this program.
*/
if (ps->speed)
ps->pread = png_create_read_struct(PNG_LIBPNG_VER_STRING, ps,
store_error, store_warning);
# ifdef PNG_USER_MEM_SUPPORTED
if (!ps->speed)
ps->pread = png_create_read_struct_2(PNG_LIBPNG_VER_STRING, ps,
store_error, store_warning, &ps->read_memory_pool, store_malloc,
store_free);
else
ps->pread = png_create_read_struct_2(PNG_LIBPNG_VER_STRING, ps,
store_error, store_warning, &ps->read_memory_pool, store_malloc,
store_free);
else
# endif
ps->pread = png_create_read_struct(PNG_LIBPNG_VER_STRING, ps, store_error,
store_warning);
if (ps->pread == NULL)
{
@@ -1945,7 +1965,7 @@ typedef struct png_modifier
/* This returns true if the test should be stopped now because it has already
* failed and it is running silently.
*/
*/
static int fail(png_modifier *pm)
{
return !pm->log && !pm->this.verbose && (pm->this.nerrors > 0 ||
@@ -2963,6 +2983,12 @@ sbit_modification_init(sbit_modification *me, png_modifier *pm, png_byte sbit)
* height of 16 rows. The width and height are stored in the FILEID and, being
* non-zero, indicate a size file.
*
* Because the PNG filter code is typically the largest CPU consumer within
* libpng itself there is a tendency to attempt to optimize it. This results in
* special case code which needs to be validated. To cause this to happen the
* 'size' images are made to use each possible filter, in so far as this is
* possible for smaller images.
*
* For palette image (colour type 3) multiple transform images are stored with
* the same bit depth to allow testing of more colour combinations -
* particularly important for testing the gamma code because libpng uses a
@@ -3121,8 +3147,10 @@ init_standard_palette(png_store *ps, png_structp pp, png_infop pi, int npalette,
for (; i<256; ++i)
tRNS[i] = 24;
if (j > 0)
png_set_tRNS(pp, pi, tRNS, j, 0/*color*/);
# ifdef PNG_WRITE_tRNS_SUPPORTED
if (j > 0)
png_set_tRNS(pp, pi, tRNS, j, 0/*color*/);
# endif
}
}
@@ -3368,6 +3396,13 @@ transform_row(png_const_structp pp, png_byte buffer[TRANSFORM_ROWMAX],
*/
#define DEPTH(bd) ((png_byte)(1U << (bd)))
/* This is just a helper for compiling on minimal systems with no write
* interlacing support.
*/
#ifndef PNG_WRITE_INTERLACING_SUPPORTED
# define png_set_interlace_handling(a) (1)
#endif
/* Make a standardized image given a an image colour type, bit depth and
* interlace type. The standard images have a very restricted range of
* rows and heights and are used for testing transforms rather than image
@@ -3521,8 +3556,12 @@ make_transform_images(png_store *ps)
{
int interlace_type;
for (interlace_type = PNG_INTERLACE_NONE;
interlace_type < PNG_INTERLACE_LAST; ++interlace_type)
# ifdef PNG_WRITE_INTERLACING_SUPPORTED
for (interlace_type = PNG_INTERLACE_NONE;
interlace_type < PNG_INTERLACE_LAST; ++interlace_type)
# else
interlace_type = PNG_INTERLACE_NONE;
# endif
{
char name[FILE_NAME_SIZE];
@@ -3678,6 +3717,9 @@ make_size_image(png_store* PNG_CONST ps, png_byte PNG_CONST colour_type,
int npasses = npasses_from_interlace_type(pp, interlace_type);
png_uint_32 y;
int pass;
# ifdef PNG_WRITE_FILTER_SUPPORTED
int nfilter = PNG_FILTER_VALUE_LAST;
# endif
png_byte image[16][SIZE_ROWMAX];
/* To help consistent error detection make the parts of this buffer
@@ -3731,7 +3773,24 @@ make_size_image(png_store* PNG_CONST ps, png_byte PNG_CONST colour_type,
continue;
}
/* Only get to here if the row has some pixels in it. */
# ifdef PNG_WRITE_FILTER_SUPPORTED
/* Only get to here if the row has some pixels in it, set the
* filters to 'all' for the very first row and thereafter to a
* single filter. It isn't well documented, but png_set_filter
* does accept a filter number (per the spec) as well as a bit
* mask.
*
* The apparent wackiness of decrementing nfilter rather than
* incrementing is so that Paeth gets used in all images bigger
* than 1 row - it's the tricky one.
*/
png_set_filter(pp, 0/*method*/,
nfilter >= PNG_FILTER_VALUE_LAST ? PNG_ALL_FILTERS : nfilter);
if (nfilter-- == 0)
nfilter = PNG_FILTER_VALUE_LAST-1;
# endif
png_write_row(pp, row);
}
}
@@ -3798,8 +3857,10 @@ make_size(png_store* PNG_CONST ps, png_byte PNG_CONST colour_type, int bdlo,
width, height, 0);
make_size_image(ps, colour_type, DEPTH(bdlo), PNG_INTERLACE_NONE,
width, height, 1);
# ifdef PNG_WRITE_INTERLACING_SUPPORTED
make_size_image(ps, colour_type, DEPTH(bdlo), PNG_INTERLACE_ADAM7,
width, height, 0);
# endif
make_size_image(ps, colour_type, DEPTH(bdlo), PNG_INTERLACE_ADAM7,
width, height, 1);
}
@@ -3994,8 +4055,12 @@ make_errors(png_modifier* PNG_CONST pm, png_byte PNG_CONST colour_type,
{
int interlace_type;
for (interlace_type = PNG_INTERLACE_NONE;
interlace_type < PNG_INTERLACE_LAST; ++interlace_type)
# ifdef PNG_WRITE_INTERLACING_SUPPORTED
for (interlace_type = PNG_INTERLACE_NONE;
interlace_type < PNG_INTERLACE_LAST; ++interlace_type)
# else
interlace_type = PNG_INTERLACE_NONE;
# endif
{
unsigned int test;
char name[FILE_NAME_SIZE];
@@ -4016,7 +4081,7 @@ make_errors(png_modifier* PNG_CONST pm, png_byte PNG_CONST colour_type,
return 1; /* keep going */
}
#endif
#endif /* PNG_WARNINGS_SUPPORTED */
static void
perform_error_test(png_modifier *pm)
@@ -10412,12 +10477,16 @@ int main(int argc, char **argv)
}
}
/* This is required because some very minimal configurations do not use it:
*/
UNUSED(fail)
return 0;
}
#else /* write not supported */
#else /* write or low level APIs not supported */
int main(void)
{
fprintf(stderr, "pngvalid: no write support in libpng, all tests skipped\n");
fprintf(stderr,
"pngvalid: no low level write support in libpng, all tests skipped\n");
/* So the test is skipped: */
return 77;
}

View File

@@ -15,7 +15,7 @@ RM=rm -f
COPY=cp
CPPFLAGS=-I. -DPNG_USER_CONFIG -DNO_GZCOMPRESS -DZ_SOLO -DNO_GZIP
CFLAGS=-O1
CFLAGS=-O1 -Wall
C=.c
O=.o

View File

@@ -15,7 +15,7 @@ RM=rm -f
COPY=cp
CPPFLAGS=-I. -DPNG_USER_CONFIG -DNO_GZCOMPRESS -DZ_SOLO -DNO_GZIP
CFLAGS=-O1
CFLAGS=-O1 -Wall
C=.c
O=.o

View File

@@ -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

View File

@@ -31,7 +31,7 @@ XLIB = -L/usr/X11R6/lib -lX11
LIBS = $(XLIB) -lm #platforms that need libm
CPPFLAGS=-I. $(XINC) -DPNG_USER_CONFIG -DNO_GZCOMPRESS -DZ_SOLO -DNO_GZIP
CFLAGS=-O1
CFLAGS=-O1 -Wall
C=.c
O=.o

View File

@@ -2,7 +2,7 @@
*
* Copyright (c) 2013 John Cunningham Bowler
*
* Last changed in libpng 1.7.0 [(PENDING RELEASE)]
* Last changed in libpng 1.6.8 [(PENDING RELEASE)]
*
* This code is released under the libpng license.
* For conditions of distribution and use, see the disclaimer
@@ -15,6 +15,7 @@
#include <stdio.h>
#include <string.h>
#include <ctype.h>
#include <limits.h>
#include <errno.h>
#include <assert.h>
@@ -49,8 +50,31 @@
# error "pngfix will not work with libpng prior to 1.6.3"
#endif
#ifdef PNG_READ_SUPPORTED
#if defined(PNG_READ_SUPPORTED) && defined(PNG_EASY_ACCESS_SUPPORTED)
/* zlib.h defines the structure z_stream, an instance of which is included
* in this structure and is required for decompressing the LZ compressed
* data in PNG files.
*/
#ifndef ZLIB_CONST
/* We must ensure that zlib uses 'const' in declarations. */
# define ZLIB_CONST
#endif
#include <zlib.h>
#ifdef const
/* zlib.h sometimes #defines const to nothing, undo this. */
# undef const
#endif
/* zlib.h has mediocre z_const use before 1.2.6, this stuff is for compatibility
* with older builds.
*/
#if ZLIB_VERNUM < 0x1260
# define PNGZ_MSG_CAST(s) png_constcast(char*,s)
# define PNGZ_INPUT_CAST(b) png_constcast(png_bytep,b)
#else
# define PNGZ_MSG_CAST(s) (s)
# define PNGZ_INPUT_CAST(b) (b)
#endif
#ifndef PNG_MAXIMUM_INFLATE_WINDOW
# error "pngfix not supported in this libpng version"
@@ -118,6 +142,11 @@
/* Is it safe to copy? */
#define SAFE_TO_COPY(chunk) (((chunk) & PNG_U32(0,0,0,32)) != 0)
/* Fix ups for builds with limited read support */
#ifndef PNG_ERROR_TEXT_SUPPORTED
# define png_error(a,b) png_err(a)
#endif
/********************************* UTILITIES **********************************/
/* UNREACHED is a value to cause an assert to fail. Because of the way the
* assert macro is written the string "UNREACHED" is produced in the error
@@ -885,10 +914,10 @@ emit_string(const char *str, FILE *out)
*/
{
for (; *str; ++str)
if (isgraph(*str))
if (isgraph(UCHAR_MAX & *str))
putc(*str, out);
else if (isspace(*str))
else if (isspace(UCHAR_MAX & *str))
putc('_', out);
else
@@ -1546,7 +1575,7 @@ chunk_end(struct chunk **chunk_var)
}
static void
chunk_init(struct chunk *chunk, struct file *file)
chunk_init(struct chunk * const chunk, struct file * const file)
/* When a chunk is initialized the file length/type/pos are copied into the
* corresponding chunk fields and the new chunk is registered in the file
* structure. There can only be one chunk at a time.
@@ -1755,7 +1784,7 @@ IDAT_end(struct IDAT **idat_var)
}
static void
IDAT_init(struct IDAT *idat, struct file *file)
IDAT_init(struct IDAT * const idat, struct file * const file)
/* When the chunk is png_IDAT instantiate an IDAT control structure in place
* of a chunk control structure. The IDAT will instantiate a chunk control
* structure using the file alloc routine.
@@ -2636,7 +2665,7 @@ zlib_check(struct file *file, png_uint_32 offset)
case ZLIB_OK:
/* Truncated stream; unrecoverable, gets converted to ZLIB_FATAL */
zlib.z.msg = png_constcast(char*, "[truncated]");
zlib.z.msg = PNGZ_MSG_CAST("[truncated]");
zlib_message(&zlib, 0/*expected*/);
/* FALL THROUGH */
@@ -2675,7 +2704,7 @@ zlib_check(struct file *file, png_uint_32 offset)
/* Output the error that wasn't output before: */
if (zlib.z.msg == NULL)
zlib.z.msg = png_constcast(char*,
zlib.z.msg = PNGZ_MSG_CAST(
"invalid distance too far back");
zlib_message(&zlib, 0/*stream error*/);
zlib_end(&zlib);
@@ -3294,6 +3323,8 @@ read_callback(png_structp png_ptr, png_bytep buffer, size_t count)
else
{
assert(chunk != NULL);
/* Set up for write, notice that repositioning the input stream
* is only necessary if something is to be read from it. Also
* notice that for the IDAT stream this must only happen once -
@@ -3307,6 +3338,8 @@ read_callback(png_structp png_ptr, png_bytep buffer, size_t count)
/* FALL THROUGH */
default:
assert(chunk != NULL);
/* NOTE: the arithmetic below overflows and gives a large positive
* png_uint_32 value until the whole chunk data has been written.
*/
@@ -3512,22 +3545,14 @@ allocate(struct file *file, int allocate_idat)
if (allocate_idat)
{
struct IDAT *idat;
assert(file->idat == NULL);
idat = &control->idat;
IDAT_init(idat, file);
file->idat = idat;
IDAT_init(&control->idat, file);
}
else /* chunk */
{
struct chunk *chunk;
assert(file->chunk == NULL);
chunk = &control->chunk;
chunk_init(chunk, file);
file->chunk = chunk;
chunk_init(&control->chunk, file);
}
}
@@ -3620,7 +3645,7 @@ read_png(struct control *control)
return rc;
}
static void
static int
one_file(struct global *global, const char *file_name, const char *out_name)
{
int rc;
@@ -3639,6 +3664,8 @@ one_file(struct global *global, const char *file_name, const char *out_name)
rc = read_png(&control);
rc |= control_end(&control);
return rc;
}
static void
@@ -4007,4 +4034,4 @@ main(void)
fprintf(stderr, "pngfix does not work without read support\n");
return 77;
}
#endif /* PNG_READ_SUPPORTED */
#endif /* PNG_READ_SUPPORTED && PNG_EASY_ACCESS_SUPPORTED */

View File

@@ -1,6 +1,6 @@
libpng-manual.txt - A description on how to use and modify libpng
libpng version 1.7.0beta19 - September 30, 2013
libpng version 1.7.0beta24 - December 15, 2013
Updated and distributed by Glenn Randers-Pehrson
<glennrp at users.sourceforge.net>
Copyright (c) 1998-2013 Glenn Randers-Pehrson
@@ -11,7 +11,7 @@ libpng-manual.txt - A description on how to use and modify libpng
Based on:
libpng versions 0.97, January 1998, through 1.7.0beta19 - September 30, 2013
libpng versions 0.97, January 1998, through 1.7.0beta24 - December 15, 2013
Updated and distributed by Glenn Randers-Pehrson
Copyright (c) 1998-2013 Glenn Randers-Pehrson
@@ -3117,6 +3117,19 @@ Here is an example of writing two private chunks, prVt and miNE:
/* Needed because miNE is not safe-to-copy */
png_set_keep_unknown_chunks(png, PNG_HANDLE_CHUNK_ALWAYS,
(png_bytep) "miNE", 1);
# if PNG_LIBPNG_VER < 10600
/* Deal with unknown chunk location bug in 1.5.x and earlier */
png_set_unknown_chunk_location(png, info, 0, PNG_HAVE_IHDR);
png_set_unknown_chunk_location(png, info, 1, PNG_AFTER_IDAT);
# endif
# if PNG_LIBPNG_VER < 10500
/* PNG_AFTER_IDAT writes two copies of the chunk prior to libpng-1.5.0,
* one before IDAT and another after IDAT, so don't use it; only use
* PNG_HAVE_IHDR location. This call resets the location previously
* set by assignment and png_set_unknown_chunk_location() for chunk 1.
*/
png_set_unknown_chunk_location(png, info, 1, PNG_HAVE_IHDR);
# endif
#endif
The high-level write interface
@@ -4608,11 +4621,11 @@ XI. Changes to Libpng from version 1.4.x to 1.5.x
From libpng-1.4.0 until 1.4.4, the png_get_uint_16 macro (but not the
function) incorrectly returned a value of type png_uint_32.
Checking for invalid palette index on read or write was added at libpng
1.5.10. When an invalid index is found, libpng issues a benign error.
This is enabled by default because this condition is an error according
to the PNG specification, Clause 11.3.2, but the error can be ignored in
each png_ptr with
Checking for invalid palette index on write was added at libpng
1.5.10. If a pixel contains an invalid (out-of-range) index libpng issues
a benign error. This is enabled by default because this condition is an
error according to the PNG specification, Clause 11.3.2, but the error can
be ignored in each png_ptr with
png_set_check_for_invalid_index(png_ptr, allowed);
@@ -5001,10 +5014,11 @@ where "rp" indicates a "restricted pointer".
Error detection in some chunks has improved; in particular the iCCP chunk
reader now does pretty complete validation of the basic format. Some bad
profiles that were previously accepted are now rejected, in particular the
very old broken Microsoft/HP sRGB profile. The PNG spec requirement that
only grayscale profiles may appear in images with color type 0 or 4 and that
even if the image only contains gray pixels, only RGB profiles may appear
profiles that were previously accepted are now accepted with a warning or
rejected, depending upon the png_set_benign_errors() setting, in particular the
very old broken Microsoft/HP 3144-byte sRGB profile. The PNG spec requirement
that only grayscale profiles may appear in images with color type 0 or 4 and
that even if the image only contains gray pixels, only RGB profiles may appear
in images with color type 2, 3, or 6, is now enforced. The sRGB chunk
is allowed to appear in images with any color type.
@@ -5013,7 +5027,9 @@ an empty language field or an empty translated keyword. Both of these
are allowed by the PNG specification, so these warnings are no longer issued.
The library now issues an error if the application attempts to set a
transform after it calls png_read_update_info().
transform after it calls png_read_update_info() or if it attempts to call
both png_read_update_info() and png_start_read_image() or to call either
of them more than once.
The default condition for benign_errors is now to treat benign errors as
warnings while reading and as errors while writing.
@@ -5034,18 +5050,19 @@ The machine-generated configure files are no longer included in branches
libpng16 and later of the GIT repository. They continue to be included
in the tarball releases, however.
Libpng-1.6.0 and later use the CMF bytes at the beginning of the IDAT stream
to set the size of the sliding window for reading instead of using the default
32-kbyte sliding window size. It was discovered that there are hundreds of PNG
files in the wild that have incorrect CMF bytes that cause libpng to now issue
a "too far back" error and reject the file. Libpng-1.6.3 provides a way to
revert to the libpng-1.5.x behavior (ignoring the CMF bytes and using a
Libpng-1.6.0 through 1.6.2 used the CMF bytes at the beginning of the IDAT
stream to set the size of the sliding window for reading instead of using the
default 32-kbyte sliding window size. It was discovered that there are
hundreds of PNG files in the wild that have incorrect CMF bytes that caused
libpng to issue a "too far back" error and reject the file. Libpng-1.6.3 and
later calculate their own safe CMF from the image dimensions, provide a way
to revert to the libpng-1.5.x behavior (ignoring the CMF bytes and using a
32-kbyte sliding window), by using
png_set_option(png_ptr, PNG_MAXIMUM_INFLATE_WINDOW,
PNG_OPTION_ON);
and provides a tool (contrib/tools/pngfix) for optimizing the CMF bytes
and provide a tool (contrib/tools/pngfix) for optimizing the CMF bytes
correctly.
Libpng-1.6.0 and libpng-1.6.1 wrote uncompressed iTXt chunks with the wrong
@@ -5231,7 +5248,8 @@ left parenthesis that follows it:
y[i] = a(x) + (int)b;
We prefer #ifdef and #ifndef to #if defined() and #if !defined()
when there is only one macro being tested.
when there is only one macro being tested. We always use parentheses
with "defined".
We prefer to express integers that are used as bit masks in hex format,
with an even number of lower-case hex digits (e.g., 0x00, 0xff, 0x0100).
@@ -5247,13 +5265,13 @@ Other rules can be inferred by inspecting the libpng source.
XVII. Y2K Compliance in libpng
September 30, 2013
December 15, 2013
Since the PNG Development group is an ad-hoc body, we can't make
an official declaration.
This is your unofficial assurance that libpng from version 0.71 and
upward through 1.7.0beta19 are Y2K compliant. It is my belief that earlier
upward through 1.7.0beta24 are Y2K compliant. It is my belief that earlier
versions were also Y2K compliant.
Libpng only has two year fields. One is a 2-byte unsigned integer

View File

@@ -1,6 +1,6 @@
.TH LIBPNG 3 "September 30, 2013"
.TH LIBPNG 3 "December 15, 2013"
.SH NAME
libpng \- Portable Network Graphics (PNG) Reference Library 1.7.0beta19
libpng \- Portable Network Graphics (PNG) Reference Library 1.7.0beta24
.SH SYNOPSIS
\fB
#include <png.h>\fP
@@ -494,7 +494,7 @@ Following is a copy of the libpng-manual.txt file that accompanies libpng.
.SH LIBPNG.TXT
libpng-manual.txt - A description on how to use and modify libpng
libpng version 1.7.0beta19 - September 30, 2013
libpng version 1.7.0beta24 - December 15, 2013
Updated and distributed by Glenn Randers-Pehrson
<glennrp at users.sourceforge.net>
Copyright (c) 1998-2013 Glenn Randers-Pehrson
@@ -505,7 +505,7 @@ libpng-manual.txt - A description on how to use and modify libpng
Based on:
libpng versions 0.97, January 1998, through 1.7.0beta19 - September 30, 2013
libpng versions 0.97, January 1998, through 1.7.0beta24 - December 15, 2013
Updated and distributed by Glenn Randers-Pehrson
Copyright (c) 1998-2013 Glenn Randers-Pehrson
@@ -3611,6 +3611,19 @@ Here is an example of writing two private chunks, prVt and miNE:
/* Needed because miNE is not safe-to-copy */
png_set_keep_unknown_chunks(png, PNG_HANDLE_CHUNK_ALWAYS,
(png_bytep) "miNE", 1);
# if PNG_LIBPNG_VER < 10600
/* Deal with unknown chunk location bug in 1.5.x and earlier */
png_set_unknown_chunk_location(png, info, 0, PNG_HAVE_IHDR);
png_set_unknown_chunk_location(png, info, 1, PNG_AFTER_IDAT);
# endif
# if PNG_LIBPNG_VER < 10500
/* PNG_AFTER_IDAT writes two copies of the chunk prior to libpng-1.5.0,
* one before IDAT and another after IDAT, so don't use it; only use
* PNG_HAVE_IHDR location. This call resets the location previously
* set by assignment and png_set_unknown_chunk_location() for chunk 1.
*/
png_set_unknown_chunk_location(png, info, 1, PNG_HAVE_IHDR);
# endif
#endif
.SS The high-level write interface
@@ -5103,11 +5116,11 @@ We removed the trailing '.' from the warning and error messages.
From libpng-1.4.0 until 1.4.4, the png_get_uint_16 macro (but not the
function) incorrectly returned a value of type png_uint_32.
Checking for invalid palette index on read or write was added at libpng
1.5.10. When an invalid index is found, libpng issues a benign error.
This is enabled by default because this condition is an error according
to the PNG specification, Clause 11.3.2, but the error can be ignored in
each png_ptr with
Checking for invalid palette index on write was added at libpng
1.5.10. If a pixel contains an invalid (out-of-range) index libpng issues
a benign error. This is enabled by default because this condition is an
error according to the PNG specification, Clause 11.3.2, but the error can
be ignored in each png_ptr with
png_set_check_for_invalid_index(png_ptr, allowed);
@@ -5496,10 +5509,11 @@ where "rp" indicates a "restricted pointer".
Error detection in some chunks has improved; in particular the iCCP chunk
reader now does pretty complete validation of the basic format. Some bad
profiles that were previously accepted are now rejected, in particular the
very old broken Microsoft/HP sRGB profile. The PNG spec requirement that
only grayscale profiles may appear in images with color type 0 or 4 and that
even if the image only contains gray pixels, only RGB profiles may appear
profiles that were previously accepted are now accepted with a warning or
rejected, depending upon the png_set_benign_errors() setting, in particular the
very old broken Microsoft/HP 3144-byte sRGB profile. The PNG spec requirement
that only grayscale profiles may appear in images with color type 0 or 4 and
that even if the image only contains gray pixels, only RGB profiles may appear
in images with color type 2, 3, or 6, is now enforced. The sRGB chunk
is allowed to appear in images with any color type.
@@ -5508,7 +5522,9 @@ an empty language field or an empty translated keyword. Both of these
are allowed by the PNG specification, so these warnings are no longer issued.
The library now issues an error if the application attempts to set a
transform after it calls png_read_update_info().
transform after it calls png_read_update_info() or if it attempts to call
both png_read_update_info() and png_start_read_image() or to call either
of them more than once.
The default condition for benign_errors is now to treat benign errors as
warnings while reading and as errors while writing.
@@ -5529,18 +5545,19 @@ The machine-generated configure files are no longer included in branches
libpng16 and later of the GIT repository. They continue to be included
in the tarball releases, however.
Libpng-1.6.0 and later use the CMF bytes at the beginning of the IDAT stream
to set the size of the sliding window for reading instead of using the default
32-kbyte sliding window size. It was discovered that there are hundreds of PNG
files in the wild that have incorrect CMF bytes that cause libpng to now issue
a "too far back" error and reject the file. Libpng-1.6.3 provides a way to
revert to the libpng-1.5.x behavior (ignoring the CMF bytes and using a
Libpng-1.6.0 through 1.6.2 used the CMF bytes at the beginning of the IDAT
stream to set the size of the sliding window for reading instead of using the
default 32-kbyte sliding window size. It was discovered that there are
hundreds of PNG files in the wild that have incorrect CMF bytes that caused
libpng to issue a "too far back" error and reject the file. Libpng-1.6.3 and
later calculate their own safe CMF from the image dimensions, provide a way
to revert to the libpng-1.5.x behavior (ignoring the CMF bytes and using a
32-kbyte sliding window), by using
png_set_option(png_ptr, PNG_MAXIMUM_INFLATE_WINDOW,
PNG_OPTION_ON);
and provides a tool (contrib/tools/pngfix) for optimizing the CMF bytes
and provide a tool (contrib/tools/pngfix) for optimizing the CMF bytes
correctly.
Libpng-1.6.0 and libpng-1.6.1 wrote uncompressed iTXt chunks with the wrong
@@ -5726,7 +5743,8 @@ left parenthesis that follows it:
y[i] = a(x) + (int)b;
We prefer #ifdef and #ifndef to #if defined() and #if !defined()
when there is only one macro being tested.
when there is only one macro being tested. We always use parentheses
with "defined".
We prefer to express integers that are used as bit masks in hex format,
with an even number of lower-case hex digits (e.g., 0x00, 0xff, 0x0100).
@@ -5742,13 +5760,13 @@ Other rules can be inferred by inspecting the libpng source.
.SH XVII. Y2K Compliance in libpng
September 30, 2013
December 15, 2013
Since the PNG Development group is an ad-hoc body, we can't make
an official declaration.
This is your unofficial assurance that libpng from version 0.71 and
upward through 1.7.0beta19 are Y2K compliant. It is my belief that earlier
upward through 1.7.0beta24 are Y2K compliant. It is my belief that earlier
versions were also Y2K compliant.
Libpng only has two year fields. One is a 2-byte unsigned integer
@@ -5955,7 +5973,7 @@ the first widely used release:
1.5.7 15 10507 15.so.15.7[.0]
1.6.0beta01-37 16 10600 16.so.16.0[.0]
1.7.0alpha01-10 17 10700 17.so.17.0[.0]
1.7.0beta01-19 17 10700 17.so.17.0[.0]
1.7.0beta01-24 17 10700 17.so.17.0[.0]
Henceforth the source version will match the shared-library minor
and patch numbers; the shared-library major version number will be
@@ -6012,7 +6030,7 @@ possible without all of you.
Thanks to Frank J. T. Wojcik for helping with the documentation.
Libpng version 1.7.0beta19 - September 30, 2013:
Libpng version 1.7.0beta24 - December 15, 2013:
Initially created in 1995 by Guy Eric Schalnat, then of Group 42, Inc.
Currently maintained by Glenn Randers-Pehrson (glennrp at users.sourceforge.net).
@@ -6035,7 +6053,7 @@ this sentence.
This code is released under the libpng license.
libpng versions 1.2.6, August 15, 2004, through 1.7.0beta19, September 30, 2013, are
libpng versions 1.2.6, August 15, 2004, through 1.7.0beta24, December 15, 2013, are
Copyright (c) 2004,2006-2007 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
@@ -6134,7 +6152,7 @@ certification mark of the Open Source Initiative.
Glenn Randers-Pehrson
glennrp at users.sourceforge.net
September 30, 2013
December 15, 2013
.\" end of man page

View File

@@ -1,6 +1,6 @@
.TH LIBPNGPF 3 "September 30, 2013"
.TH LIBPNGPF 3 "December 15, 2013"
.SH NAME
libpng \- Portable Network Graphics (PNG) Reference Library 1.7.0beta19
libpng \- Portable Network Graphics (PNG) Reference Library 1.7.0beta24
(private functions)
.SH SYNOPSIS
\fB#include \fI"pngpriv.h"

2
png.5
View File

@@ -1,4 +1,4 @@
.TH PNG 5 "September 30, 2013"
.TH PNG 5 "December 15, 2013"
.SH NAME
png \- Portable Network Graphics (PNG) format
.SH DESCRIPTION

75
png.c
View File

@@ -14,7 +14,7 @@
#include "pngpriv.h"
/* Generate a compiler error if there is an old png.h in the search path. */
typedef png_libpng_version_1_7_0beta19 Your_png_h_is_not_version_1_7_0beta19;
typedef png_libpng_version_1_7_0beta24 Your_png_h_is_not_version_1_7_0beta24;
/* Tells libpng that we have already handled the first "num_bytes" bytes
* of the PNG file signature. If the PNG data is embedded into another
@@ -201,6 +201,7 @@ png_user_version_check(png_structrp png_ptr, png_const_charp user_png_ver)
pos = png_safecat(m, (sizeof m), pos, user_png_ver);
pos = png_safecat(m, (sizeof m), pos, " but running with ");
pos = png_safecat(m, (sizeof m), pos, png_libpng_ver);
PNG_UNUSED(pos)
png_warning(png_ptr, m);
#endif
@@ -258,6 +259,10 @@ png_create_png_struct,(png_const_charp user_png_ver, png_voidp error_ptr,
*/
# ifdef PNG_USER_MEM_SUPPORTED
png_set_mem_fn(&create_struct, mem_ptr, malloc_fn, free_fn);
# else
PNG_UNUSED(mem_ptr)
PNG_UNUSED(malloc_fn)
PNG_UNUSED(free_fn)
# endif
/* (*error_fn) can return control to the caller after the error_ptr is set,
@@ -691,13 +696,13 @@ png_get_copyright(png_const_structrp png_ptr)
#else
# ifdef __STDC__
return PNG_STRING_NEWLINE \
"libpng version 1.7.0beta19 - September 30, 2013" PNG_STRING_NEWLINE \
"libpng version 1.7.0beta24 - December 15, 2013" PNG_STRING_NEWLINE \
"Copyright (c) 1998-2013 Glenn Randers-Pehrson" PNG_STRING_NEWLINE \
"Copyright (c) 1996-1997 Andreas Dilger" PNG_STRING_NEWLINE \
"Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc." \
PNG_STRING_NEWLINE;
# else
return "libpng version 1.7.0beta19 - September 30, 2013\
return "libpng version 1.7.0beta24 - December 15, 2013\
Copyright (c) 1998-2013 Glenn Randers-Pehrson\
Copyright (c) 1996-1997 Andreas Dilger\
Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.";
@@ -744,6 +749,63 @@ png_get_header_version(png_const_structrp png_ptr)
#endif
}
#ifdef PNG_BUILD_GRAYSCALE_PALETTE_SUPPORTED
/* NOTE: this routine is not used internally! */
/* Build a grayscale palette. Palette is assumed to be 1 << bit_depth
* large of png_color. This lets grayscale images be treated as
* paletted. Most useful for gamma correction and simplification
* of code. This API is not used internally.
*/
void PNGAPI
png_build_grayscale_palette(int bit_depth, png_colorp palette)
{
int num_palette;
int color_inc;
int i;
int v;
png_debug(1, "in png_do_build_grayscale_palette");
if (palette == NULL)
return;
switch (bit_depth)
{
case 1:
num_palette = 2;
color_inc = 0xff;
break;
case 2:
num_palette = 4;
color_inc = 0x55;
break;
case 4:
num_palette = 16;
color_inc = 0x11;
break;
case 8:
num_palette = 256;
color_inc = 1;
break;
default:
num_palette = 0;
color_inc = 0;
break;
}
for (i = 0, v = 0; i < num_palette; i++, v += color_inc)
{
palette[i].red = (png_byte)v;
palette[i].green = (png_byte)v;
palette[i].blue = (png_byte)v;
}
}
#endif
#ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED
int PNGAPI
png_handle_as_unknown(png_const_structrp png_ptr, png_const_bytep chunk_name)
@@ -1630,6 +1692,7 @@ png_icc_profile_error(png_const_structrp png_ptr, png_colorspacerp colorspace,
# endif
/* The 'reason' is an arbitrary message, allow +79 maximum 195 */
pos = png_safecat(message, (sizeof message), pos, reason);
PNG_UNUSED(pos)
/* This is recoverable, but make it unconditionally an app_error on write to
* avoid writing invalid ICC profiles into PNG files. (I.e. we handle them
@@ -3008,10 +3071,14 @@ png_fixed(png_const_structrp png_ptr, double fp, png_const_charp text)
/* Overflow */
png_fixed_error(png_ptr, text);
# ifndef PNG_ERROR_TEXT_SUPPORTED
PNG_UNUSED(text)
# endif
}
#endif
#if defined(PNG_READ_GAMMA_SUPPORTED) || \
#if defined(PNG_GAMMA_SUPPORTED) || defined(PNG_COLORSPACE_SUPPORTED) ||\
defined(PNG_INCH_CONVERSIONS_SUPPORTED) || defined(PNG_READ_pHYs_SUPPORTED)
/* muldiv functions */
/* This API takes signed arguments and rounds the result to the nearest

20
png.h
View File

@@ -1,7 +1,7 @@
/* png.h - header file for PNG reference library
*
* libpng version 1.7.0beta19 - September 30, 2013
* libpng version 1.7.0beta24 - December 15, 2013
* Copyright (c) 1998-2013 Glenn Randers-Pehrson
* (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
* (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
@@ -11,7 +11,7 @@
* Authors and maintainers:
* libpng versions 0.71, May 1995, through 0.88, January 1996: Guy Schalnat
* libpng versions 0.89c, June 1996, through 0.96, May 1997: Andreas Dilger
* libpng versions 0.97, January 1998, through 1.7.0beta19 - September 30, 2013: Glenn
* libpng versions 0.97, January 1998, through 1.7.0beta24 - December 15, 2013: Glenn
* See also "Contributing Authors", below.
*
* Note about libpng version numbers:
@@ -168,7 +168,7 @@
* 1.5.7 15 10507 15.so.15.7[.0]
* 1.6.0beta01-37 16 10600 16.so.16.0[.0]
* 1.7.0alpha01-10 17 10700 17.so.17.0[.0]
* 1.7.0beta01-19 17 10700 17.so.17.0[.0]
* 1.7.0beta01-24 17 10700 17.so.17.0[.0]
*
* Henceforth the source version will match the shared-library major
* and minor numbers; the shared-library major version number will be
@@ -200,7 +200,7 @@
*
* This code is released under the libpng license.
*
* libpng versions 1.2.6, August 15, 2004, through 1.7.0beta19, September 30, 2013, are
* libpng versions 1.2.6, August 15, 2004, through 1.7.0beta24, December 15, 2013, are
* Copyright (c) 2004, 2006-2013 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:
@@ -312,13 +312,13 @@
* Y2K compliance in libpng:
* =========================
*
* September 30, 2013
* December 15, 2013
*
* Since the PNG Development group is an ad-hoc body, we can't make
* an official declaration.
*
* This is your unofficial assurance that libpng from version 0.71 and
* upward through 1.7.0beta19 are Y2K compliant. It is my belief that
* upward through 1.7.0beta24 are Y2K compliant. It is my belief that
* earlier versions were also Y2K compliant.
*
* Libpng only has two year fields. One is a 2-byte unsigned integer
@@ -378,9 +378,9 @@
*/
/* Version information for png.h - this should match the version in png.c */
#define PNG_LIBPNG_VER_STRING "1.7.0beta19"
#define PNG_LIBPNG_VER_STRING "1.7.0beta24"
#define PNG_HEADER_VERSION_STRING \
" libpng version 1.7.0beta19 - September 30, 2013\n"
" libpng version 1.7.0beta24 - December 15, 2013\n"
#define PNG_LIBPNG_VER_SONUM 17
#define PNG_LIBPNG_VER_DLLNUM 17
@@ -394,7 +394,7 @@
* PNG_LIBPNG_VER_STRING, omitting any leading zero:
*/
#define PNG_LIBPNG_VER_BUILD 19
#define PNG_LIBPNG_VER_BUILD 24
/* Release Status */
#define PNG_LIBPNG_BUILD_ALPHA 1
@@ -640,7 +640,7 @@ extern "C" {
/* This triggers a compiler error in png.c, if png.c and png.h
* do not agree upon the version number.
*/
typedef char* png_libpng_version_1_7_0beta19;
typedef char* png_libpng_version_1_7_0beta24;
/* Basic control structions. Read libpng-manual.txt or libpng.3 for more info.
*

View File

@@ -1,7 +1,7 @@
/* pngconf.h - machine configurable file for libpng
*
* libpng version 1.7.0beta19 - September 30, 2013
* libpng version 1.7.0beta24 - December 15, 2013
*
* Copyright (c) 1998-2013 Glenn Randers-Pehrson
* (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)

View File

@@ -1,11 +1,11 @@
/* pngdebug.h - Debugging macros for libpng, also used in pngtest.c
*
* Copyright (c) 1998-2011 Glenn Randers-Pehrson
* Copyright (c) 1998-2013 Glenn Randers-Pehrson
* (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
* (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
*
* Last changed in libpng 1.5.0 [January 6, 2011]
* Last changed in libpng 1.5.18 [(PENDING RELEASE)]
*
* This code is released under the libpng license.
* For conditions of distribution and use, see the disclaimer
@@ -25,7 +25,7 @@
* (actually ((void)0)).
*
* level: level of detail of message, starting at 0. A level 'n'
* message is preceded by 'n' tab characters (not implemented
* message is preceded by 'n' 3-space indentations (not implemented
* on Microsoft compilers unless PNG_DEBUG_FILE is also
* defined, to allow debug DLL compilation with no standard IO).
* message: a printf(3) style text string. A trailing '\n' is added
@@ -77,32 +77,29 @@
# endif /* PNG_DEBUG_FILE */
# if (PNG_DEBUG > 1)
/* Note: ["%s"m PNG_STRING_NEWLINE] probably does not work on
* non-ISO compilers
*/
# ifdef __STDC__
# ifndef png_debug
# define png_debug(l,m) \
do { \
int num_tabs=l; \
fprintf(PNG_DEBUG_FILE,"%s"m PNG_STRING_NEWLINE,(num_tabs==1 ? "\t" : \
(num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":"")))); \
fprintf(PNG_DEBUG_FILE,"%s" m PNG_STRING_NEWLINE,(num_tabs==1 ? " " : \
(num_tabs==2 ? " " : (num_tabs>2 ? " " : "")))); \
} while (0)
# endif
# ifndef png_debug1
# define png_debug1(l,m,p1) \
do { \
int num_tabs=l; \
fprintf(PNG_DEBUG_FILE,"%s"m PNG_STRING_NEWLINE,(num_tabs==1 ? "\t" : \
(num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))),p1); \
fprintf(PNG_DEBUG_FILE,"%s" m PNG_STRING_NEWLINE,(num_tabs==1 ? " " : \
(num_tabs==2 ? " " : (num_tabs>2 ? " " : ""))),p1); \
} while (0)
# endif
# ifndef png_debug2
# define png_debug2(l,m,p1,p2) \
do { \
int num_tabs=l; \
fprintf(PNG_DEBUG_FILE,"%s"m PNG_STRING_NEWLINE,(num_tabs==1 ? "\t" : \
(num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))),p1,p2); \
fprintf(PNG_DEBUG_FILE,"%s" m PNG_STRING_NEWLINE,(num_tabs==1 ? " " : \
(num_tabs==2 ? " " : (num_tabs>2 ? " " : ""))),p1,p2);\
} while (0)
# endif
# else /* __STDC __ */

View File

@@ -382,6 +382,10 @@ png_benign_error(png_const_structrp png_ptr, png_const_charp error_message)
# endif
png_error(png_ptr, error_message);
}
# ifndef PNG_ERROR_TEXT_SUPPORTED
PNG_UNUSED(error_message)
# endif
}
void /* PRIVATE */
@@ -391,6 +395,10 @@ png_app_warning(png_const_structrp png_ptr, png_const_charp error_message)
png_warning(png_ptr, error_message);
else
png_error(png_ptr, error_message);
# ifndef PNG_ERROR_TEXT_SUPPORTED
PNG_UNUSED(error_message)
# endif
}
void /* PRIVATE */
@@ -400,6 +408,10 @@ png_app_error(png_const_structrp png_ptr, png_const_charp error_message)
png_warning(png_ptr, error_message);
else
png_error(png_ptr, error_message);
# ifndef PNG_ERROR_TEXT_SUPPORTED
PNG_UNUSED(error_message)
# endif
}
#endif /* BENIGN_ERRORS */
@@ -416,7 +428,8 @@ static PNG_CONST char png_digit[16] = {
};
#define PNG_MAX_ERROR_TEXT 196 /* Currently limited be profile_error in png.c */
#if defined(PNG_WARNINGS_SUPPORTED) || defined(PNG_ERROR_TEXT_SUPPORTED)
#if defined(PNG_WARNINGS_SUPPORTED) || \
(defined(PNG_READ_SUPPORTED) && defined(PNG_ERROR_TEXT_SUPPORTED))
static void /* PRIVATE */
png_format_buffer(png_const_structrp png_ptr, png_charp buffer, png_const_charp
error_message)
@@ -506,6 +519,10 @@ png_chunk_benign_error(png_const_structrp png_ptr, png_const_charp
else
png_chunk_error(png_ptr, error_message);
# ifndef PNG_ERROR_TEXT_SUPPORTED
PNG_UNUSED(error_message)
# endif
}
#endif
#endif /* PNG_READ_SUPPORTED */
@@ -513,6 +530,10 @@ png_chunk_benign_error(png_const_structrp png_ptr, png_const_charp
void /* PRIVATE */
png_chunk_report(png_const_structrp png_ptr, png_const_charp message, int error)
{
# ifndef PNG_WARNINGS_SUPPORTED
PNG_UNUSED(message)
# endif
/* This is always supported, but for just read or just write it
* unconditionally does the right thing.
*/

View File

@@ -73,9 +73,10 @@ png_malloc_base,(png_const_structrp png_ptr, png_alloc_size_t size),
* to implement a user memory handler. This checks to be sure it isn't
* called with big numbers.
*/
#ifdef PNG_USER_MEM_SUPPORTED
#ifndef PNG_USER_MEM_SUPPORTED
PNG_UNUSED(png_ptr)
#endif
if (size > 0 && size <= PNG_SIZE_MAX
# ifdef PNG_MAX_MALLOC_64K
&& size <= 65536U
@@ -95,6 +96,8 @@ png_malloc_base,(png_const_structrp png_ptr, png_alloc_size_t size),
return NULL;
}
#if defined(PNG_TEXT_SUPPORTED) || defined(PNG_sPLT_SUPPORTED) ||\
defined(PNG_STORE_UNKNOWN_CHUNKS_SUPPORTED)
/* This is really here only to work round a spurious warning in GCC 4.6 and 4.7
* that arises because of the checks in png_realloc_array that are repeated in
* png_malloc_array.
@@ -165,6 +168,7 @@ png_realloc_array,(png_structrp png_ptr, png_const_voidp old_array,
return NULL; /* error */
}
#endif /* TEXT || sPLT || STORE_UNKNOWN_CHUNKS */
/* Various functions that have different error handling are derived from this.
* png_malloc always exists, but if PNG_USER_MEM_SUPPORTED is defined a separate

View File

@@ -185,7 +185,7 @@ void /* PRIVATE */
png_push_read_chunk(png_structrp png_ptr, png_inforp info_ptr)
{
png_uint_32 chunk_name;
#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
#ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED
int keep; /* unknown handling method */
#endif
@@ -271,7 +271,7 @@ png_push_read_chunk(png_structrp png_ptr, png_inforp info_ptr)
png_push_have_end(png_ptr, info_ptr);
}
#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
#ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED
else if ((keep = png_chunk_unknown_handling(png_ptr, chunk_name)) != 0)
{
if (png_ptr->push_length + 4 > png_ptr->buffer_size)
@@ -285,8 +285,8 @@ png_push_read_chunk(png_structrp png_ptr, png_inforp info_ptr)
if (chunk_name == png_PLTE)
png_ptr->mode |= PNG_HAVE_PLTE;
}
#endif
else if (chunk_name == png_PLTE)
{
if (png_ptr->push_length + 4 > png_ptr->buffer_size)
@@ -528,8 +528,8 @@ png_push_read_chunk(png_structrp png_ptr, png_inforp info_ptr)
png_handle_iTXt(png_ptr, info_ptr, png_ptr->push_length);
}
#endif
else
{
if (png_ptr->push_length + 4 > png_ptr->buffer_size)

176
pngpriv.h
View File

@@ -149,9 +149,15 @@
* typically the target FPU. If the FPU has been set to NEON (-mfpu=neon
* with GCC) then the compiler will define __ARM_NEON__ and we can rely
* unconditionally on NEON instructions not crashing, otherwise we must
* disable use of NEON instructions:
* disable use of NEON instructions.
*
* NOTE: at present these optimizations depend on 'ALIGNED_MEMORY', so they
* can only be turned on automatically if that is supported too. If
* PNG_ARM_NEON_OPT is set in CPPFLAGS (to >0) then arm/arm_init.c will fail
* to compile with an appropriate #error if ALIGNED_MEMORY has been turned
* off.
*/
# ifdef __ARM_NEON__
# if defined(__ARM_NEON__) && defined(PNG_ALIGNED_MEMORY_SUPPORTED)
# define PNG_ARM_NEON_OPT 2
# else
# define PNG_ARM_NEON_OPT 0
@@ -163,7 +169,49 @@
* callbacks to do this.
*/
# define PNG_FILTER_OPTIMIZATIONS png_init_filter_functions_neon
#endif
/* By default the 'intrinsics' code in arm/filter_neon_intrinsics.c is used
* if possible - if __ARM_NEON__ is set and the compiler version is not known
* to be broken. This is control by PNG_ARM_NEON_IMPLEMENTATION which can
* be:
*
* 1 The intrinsics code (the default with __ARM_NEON__)
* 2 The hand coded assembler (the default without __ARM_NEON__)
*
* It is possible to set PNG_ARM_NEON_IMPLEMENTATION in CPPFLAGS, however
* this is *NOT* supported and may cease to work even after a minor revision
* to libpng. It *is* valid to do this for testing purposes, e.g. speed
* testing or a new compiler, but the results should be communicated to the
* libpng implementation list for incorporation in the next minor release.
*/
# ifndef PNG_ARM_NEON_IMPLEMENTATION
# ifdef __ARM_NEON__
# if defined(__clang__)
/* At present it is unknown by the libpng developers which versions
* of clang support the intrinsics, however some or perhaps all
* versions do not work with the assembler so this may be
* irrelevant, so just use the default (do nothing here.)
*/
# elif defined(__GNUC__)
/* GCC 4.5.4 NEON support is known to be broken. 4.6.3 is known to
* work, so if this *is* GCC, or G++, look for a version >4.5
*/
# if __GNUC__ < 4 || (__GNUC__ == 4 && __GNUC_MINOR__ < 6)
# define PNG_ARM_NEON_IMPLEMENTATION 2
# endif /* no GNUC support */
# endif /* __GNUC__ */
# else /* !defined __ARM_NEON__ */
/* The 'intrinsics' code simply won't compile without this -mfpu=neon:
*/
# define PNG_ARM_NEON_IMPLEMENTATION 2
# endif /* __ARM_NEON__ */
# endif /* !defined PNG_ARM_NEON_IMPLEMENTATION */
# ifndef PNG_ARM_NEON_IMPLEMENTATION
/* Use the intrinsics code by default. */
# define PNG_ARM_NEON_IMPLEMENTATION 1
# endif
#endif /* PNG_ARM_NEON_OPT > 0 */
/* Is this a build of a DLL where compilation of the object modules requires
* different preprocessor settings to those required for a simple library? If
@@ -1088,7 +1136,7 @@ PNG_INTERNAL_FUNCTION(void,png_read_finish_IDAT,(png_structrp png_ptr),
PNG_INTERNAL_FUNCTION(void,png_read_finish_row,(png_structrp png_ptr),
PNG_EMPTY);
/* Finish a row while reading, dealing with interlacing passes, etc. */
#endif
#endif /* PNG_SEQUENTIAL_READ_SUPPORTED */
/* Initialize the row buffers, etc. */
PNG_INTERNAL_FUNCTION(void,png_read_start_row,(png_structrp png_ptr),PNG_EMPTY);
@@ -1099,32 +1147,7 @@ PNG_INTERNAL_FUNCTION(void,png_read_transform_info,(png_structrp png_ptr,
png_inforp info_ptr),PNG_EMPTY);
#endif
/* These are the functions that do the transformations */
#ifdef PNG_READ_FILLER_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_read_filler,(png_row_infop row_info,
png_bytep row, png_uint_32 filler, png_uint_32 flags),PNG_EMPTY);
#endif
#ifdef PNG_READ_SWAP_ALPHA_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_read_swap_alpha,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_WRITE_SWAP_ALPHA_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_write_swap_alpha,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_INVERT_ALPHA_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_read_invert_alpha,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_WRITE_INVERT_ALPHA_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_write_invert_alpha,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
/* Shared transform functions, defined in pngtran.c */
#if defined(PNG_WRITE_FILLER_SUPPORTED) || \
defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
PNG_INTERNAL_FUNCTION(void,png_do_strip_channel,(png_row_infop row_info,
@@ -1144,96 +1167,16 @@ PNG_INTERNAL_FUNCTION(void,png_do_packswap,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_RGB_TO_GRAY_SUPPORTED
PNG_INTERNAL_FUNCTION(int,png_do_rgb_to_gray,(png_structrp png_ptr,
png_row_infop row_info, png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_GRAY_TO_RGB_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_gray_to_rgb,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_PACK_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_unpack,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_SHIFT_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_unshift,(png_row_infop row_info,
png_bytep row, png_const_color_8p sig_bits),PNG_EMPTY);
#endif
#if defined(PNG_READ_INVERT_SUPPORTED) || defined(PNG_WRITE_INVERT_SUPPORTED)
PNG_INTERNAL_FUNCTION(void,png_do_invert,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_scale_16_to_8,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_STRIP_16_TO_8_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_chop,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_READ_QUANTIZE_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_quantize,(png_row_infop row_info,
png_bytep row, png_const_bytep palette_lookup,
png_const_bytep quantize_lookup),PNG_EMPTY);
# ifdef PNG_CORRECT_PALETTE_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_correct_palette,(png_structrp png_ptr,
png_colorp palette, int num_palette),PNG_EMPTY);
# endif
#endif
#if defined(PNG_READ_BGR_SUPPORTED) || defined(PNG_WRITE_BGR_SUPPORTED)
PNG_INTERNAL_FUNCTION(void,png_do_bgr,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
#ifdef PNG_WRITE_PACK_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_pack,(png_row_infop row_info,
png_bytep row, png_uint_32 bit_depth),PNG_EMPTY);
#endif
#ifdef PNG_WRITE_SHIFT_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_shift,(png_row_infop row_info,
png_bytep row, png_const_color_8p bit_depth),PNG_EMPTY);
#endif
#if defined(PNG_READ_BACKGROUND_SUPPORTED) ||\
defined(PNG_READ_ALPHA_MODE_SUPPORTED)
PNG_INTERNAL_FUNCTION(void,png_do_compose,(png_row_infop row_info,
png_bytep row, png_structrp png_ptr),PNG_EMPTY);
#endif
#ifdef PNG_READ_GAMMA_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_gamma,(png_row_infop row_info,
png_bytep row, png_structrp png_ptr),PNG_EMPTY);
#endif
#ifdef PNG_READ_ALPHA_MODE_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_encode_alpha,(png_row_infop row_info,
png_bytep row, png_structrp png_ptr),PNG_EMPTY);
#endif
#ifdef PNG_READ_EXPAND_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_expand_palette,(png_row_infop row_info,
png_bytep row, png_const_colorp palette, png_const_bytep trans,
int num_trans),PNG_EMPTY);
PNG_INTERNAL_FUNCTION(void,png_do_expand,(png_row_infop row_info,
png_bytep row, png_const_color_16p trans_color),PNG_EMPTY);
#endif
#ifdef PNG_READ_EXPAND_16_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_expand_16,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
/* The following decodes the appropriate chunks, and does error correction,
* then calls the appropriate callback for the chunk if it is valid.
*/
@@ -1334,7 +1277,6 @@ PNG_INTERNAL_FUNCTION(void,png_handle_zTXt,(png_structrp png_ptr,
PNG_INTERNAL_FUNCTION(void,png_check_chunk_name,(png_structrp png_ptr,
png_uint_32 chunk_name),PNG_EMPTY);
#ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_handle_unknown,(png_structrp png_ptr,
png_inforp info_ptr, png_uint_32 length, int keep),PNG_EMPTY);
/* This is the function that gets called for unknown chunks. The 'keep'
@@ -1344,14 +1286,13 @@ PNG_INTERNAL_FUNCTION(void,png_handle_unknown,(png_structrp png_ptr,
*/
#if defined(PNG_READ_UNKNOWN_CHUNKS_SUPPORTED) ||\
defined(PNG_HANDLE_AS_UNKNOWN_SUPPORTED)
defined(PNG_HANDLE_AS_UNKNOWN_SUPPORTED)
PNG_INTERNAL_FUNCTION(int,png_chunk_unknown_handling,
(png_const_structrp png_ptr, png_uint_32 chunk_name),PNG_EMPTY);
/* Exactly as the API png_handle_as_unknown() except that the argument is a
* 32-bit chunk name, not a string.
*/
#endif /* READ_UNKNOWN_CHUNKS || HANDLE_AS_UNKNOWN */
#endif /* PNG_SET_UNKNOWN_CHUNKS_SUPPORTED */
/* Handle the transformations for reading and writing */
#ifdef PNG_READ_TRANSFORMS_SUPPORTED
@@ -1422,13 +1363,6 @@ PNG_INTERNAL_FUNCTION(void,png_push_read_iTXt,(png_structrp png_ptr,
#endif /* PNG_PROGRESSIVE_READ_SUPPORTED */
#ifdef PNG_MNG_FEATURES_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_do_read_intrapixel,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
PNG_INTERNAL_FUNCTION(void,png_do_write_intrapixel,(png_row_infop row_info,
png_bytep row),PNG_EMPTY);
#endif
/* Added at libpng version 1.6.0 */
#ifdef PNG_GAMMA_SUPPORTED
PNG_INTERNAL_FUNCTION(void,png_colorspace_set_gamma,(png_const_structrp png_ptr,
@@ -1770,7 +1704,7 @@ PNG_INTERNAL_FUNCTION(int,png_check_fp_string,(png_const_charp string,
png_size_t size),PNG_EMPTY);
#endif /* pCAL || sCAL */
#if defined(PNG_READ_GAMMA_SUPPORTED) ||\
#if defined(PNG_GAMMA_SUPPORTED) ||\
defined(PNG_INCH_CONVERSIONS_SUPPORTED) || defined(PNG_READ_pHYs_SUPPORTED)
/* Added at libpng version 1.5.0 */
/* This is a utility to provide a*times/div (rounded) and indicate

271
pngread.c
View File

@@ -305,6 +305,72 @@ png_start_read_image(png_structrp png_ptr)
#endif /* PNG_SEQUENTIAL_READ_SUPPORTED */
#ifdef PNG_SEQUENTIAL_READ_SUPPORTED
#ifdef PNG_MNG_FEATURES_SUPPORTED
/* Undoes intrapixel differencing,
* NOTE: this is apparently only supported in the 'sequential' reader.
*/
static void
png_do_read_intrapixel(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_read_intrapixel");
if (
(row_info->color_type & PNG_COLOR_MASK_COLOR))
{
int bytes_per_pixel;
png_uint_32 row_width = row_info->width;
if (row_info->bit_depth == 8)
{
png_bytep rp;
png_uint_32 i;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 3;
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 4;
else
return;
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
*(rp) = (png_byte)((256 + *rp + *(rp + 1)) & 0xff);
*(rp+2) = (png_byte)((256 + *(rp + 2) + *(rp + 1)) & 0xff);
}
}
else if (row_info->bit_depth == 16)
{
png_bytep rp;
png_uint_32 i;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 6;
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 8;
else
return;
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
png_uint_32 s0 = (*(rp ) << 8) | *(rp + 1);
png_uint_32 s1 = (*(rp + 2) << 8) | *(rp + 3);
png_uint_32 s2 = (*(rp + 4) << 8) | *(rp + 5);
png_uint_32 red = (s0 + s1 + 65536) & 0xffff;
png_uint_32 blue = (s2 + s1 + 65536) & 0xffff;
*(rp ) = (png_byte)((red >> 8) & 0xff);
*(rp + 1) = (png_byte)(red & 0xff);
*(rp + 4) = (png_byte)((blue >> 8) & 0xff);
*(rp + 5) = (png_byte)(blue & 0xff);
}
}
}
}
#endif /* PNG_MNG_FEATURES_SUPPORTED */
void PNGAPI
png_read_row(png_structrp png_ptr, png_bytep row, png_bytep dsp_row)
{
@@ -489,7 +555,6 @@ png_read_row(png_structrp png_ptr, png_bytep row, png_bytep dsp_row)
}
#endif
#ifdef PNG_READ_TRANSFORMS_SUPPORTED
if (png_ptr->transformations)
png_do_read_transformations(png_ptr, &row_info);
@@ -1125,12 +1190,11 @@ png_read_png(png_structrp png_ptr, png_inforp info_ptr,
/* Arguments to png_image_finish_read: */
/* Encoding of PNG data (used by the color-map code) */
/* TODO: change these, dang, ANSI-C reserves the 'E' namespace. */
# define E_NOTSET 0 /* File encoding not yet known */
# define E_sRGB 1 /* 8-bit encoded to sRGB gamma */
# define E_LINEAR 2 /* 16-bit linear: not encoded, NOT pre-multiplied! */
# define E_FILE 3 /* 8-bit encoded to file gamma, not sRGB or linear */
# define E_LINEAR8 4 /* 8-bit linear: only from a file value */
# define P_NOTSET 0 /* File encoding not yet known */
# define P_sRGB 1 /* 8-bit encoded to sRGB gamma */
# define P_LINEAR 2 /* 16-bit linear: not encoded, NOT pre-multiplied! */
# define P_FILE 3 /* 8-bit encoded to file gamma, not sRGB or linear */
# define P_LINEAR8 4 /* 8-bit linear: only from a file value */
/* Color-map processing: after libpng has run on the PNG image further
* processing may be needed to conver the data to color-map indicies.
@@ -1161,7 +1225,7 @@ typedef struct
png_voidp first_row;
ptrdiff_t row_bytes; /* step between rows */
int file_encoding; /* E_ values above */
png_fixed_point gamma_to_linear; /* For E_FILE, reciprocal of gamma */
png_fixed_point gamma_to_linear; /* For P_FILE, reciprocal of gamma */
int colormap_processing; /* PNG_CMAP_ values above */
} png_image_read_control;
@@ -1293,7 +1357,7 @@ png_image_read_header(png_voidp argument)
#ifdef PNG_COLORSPACE_SUPPORTED
/* Does the colorspace match sRGB? If there is no color endpoint
* (colorant) information assume yes, otherwise require the
* 'ENDPOINTS_MATCHE_sRGB' colorspace flag to have been set. If the
* 'ENDPOINTS_MATCHP_sRGB' colorspace flag to have been set. If the
* colorspace has been determined to be invalid ignore it.
*/
if ((format & PNG_FORMAT_FLAG_COLOR) != 0 && ((png_ptr->colorspace.flags
@@ -1482,17 +1546,24 @@ png_image_skip_unused_chunks(png_structrp png_ptr)
*
* Or image data handling:
*
* tRNS, bKGD, gAMA, cHRM, sRGB, iCCP and sBIT.
* tRNS, bKGD, gAMA, cHRM, sRGB, [iCCP] and sBIT.
*
* This provides a small performance improvement and eliminates any
* potential vulnerability to security problems in the unused chunks.
*
* At present the iCCP chunk data isn't used, so iCCP chunk can be ignored
* too. This allows the simplified API to be compiled without iCCP support,
* however if the support is there the chunk is still checked to detect
* errors (which are unfortunately quite common.)
*/
{
static PNG_CONST png_byte chunks_to_process[] = {
98, 75, 71, 68, '\0', /* bKGD */
99, 72, 82, 77, '\0', /* cHRM */
103, 65, 77, 65, '\0', /* gAMA */
# ifdef PNG_READ_iCCP_SUPPORTED
105, 67, 67, 80, '\0', /* iCCP */
# endif
115, 66, 73, 84, '\0', /* sBIT */
115, 82, 71, 66, '\0', /* sRGB */
};
@@ -1529,25 +1600,25 @@ set_file_encoding(png_image_read_control *display)
{
if (png_gamma_not_sRGB(g))
{
display->file_encoding = E_FILE;
display->file_encoding = P_FILE;
display->gamma_to_linear = png_reciprocal(g);
}
else
display->file_encoding = E_sRGB;
display->file_encoding = P_sRGB;
}
else
display->file_encoding = E_LINEAR8;
display->file_encoding = P_LINEAR8;
}
static unsigned int
decode_gamma(png_image_read_control *display, png_uint_32 value, int encoding)
{
if (encoding == E_FILE) /* double check */
if (encoding == P_FILE) /* double check */
encoding = display->file_encoding;
if (encoding == E_NOTSET) /* must be the file encoding */
if (encoding == P_NOTSET) /* must be the file encoding */
{
set_file_encoding(display);
encoding = display->file_encoding;
@@ -1555,18 +1626,18 @@ decode_gamma(png_image_read_control *display, png_uint_32 value, int encoding)
switch (encoding)
{
case E_FILE:
case P_FILE:
value = png_gamma_16bit_correct(value*257, display->gamma_to_linear);
break;
case E_sRGB:
case P_sRGB:
value = png_sRGB_table[value];
break;
case E_LINEAR:
case P_LINEAR:
break;
case E_LINEAR8:
case P_LINEAR8:
value *= 257;
break;
@@ -1585,9 +1656,9 @@ png_colormap_compose(png_image_read_control *display,
png_uint_32 background, int encoding)
{
/* The file value is composed on the background, the background has the given
* encoding and so does the result, the file is encoded with E_FILE and the
* encoding and so does the result, the file is encoded with P_FILE and the
* file and alpha are 8-bit values. The (output) encoding will always be
* E_LINEAR or E_sRGB.
* P_LINEAR or P_sRGB.
*/
png_uint_32 f = decode_gamma(display, foreground, foreground_encoding);
png_uint_32 b = decode_gamma(display, background, encoding);
@@ -1597,7 +1668,7 @@ png_colormap_compose(png_image_read_control *display,
*/
f = f * alpha + b * (255-alpha);
if (encoding == E_LINEAR)
if (encoding == P_LINEAR)
{
/* Scale to 65535; divide by 255, approximately (in fact this is extremely
* accurate, it divides by 255.00000005937181414556, with no overflow.)
@@ -1607,13 +1678,13 @@ png_colormap_compose(png_image_read_control *display,
f = (f+32768) >> 16;
}
else /* E_sRGB */
else /* P_sRGB */
f = PNG_sRGB_FROM_LINEAR(f);
return f;
}
/* NOTE: E_LINEAR values to this routine must be 16-bit, but E_FILE values must
/* NOTE: P_LINEAR values to this routine must be 16-bit, but P_FILE values must
* be 8-bit.
*/
static void
@@ -1623,7 +1694,7 @@ png_create_colormap_entry(png_image_read_control *display,
{
png_imagep image = display->image;
const int output_encoding = (image->format & PNG_FORMAT_FLAG_LINEAR) ?
E_LINEAR : E_sRGB;
P_LINEAR : P_sRGB;
const int convert_to_Y = (image->format & PNG_FORMAT_FLAG_COLOR) == 0 &&
(red != green || green != blue);
@@ -1633,18 +1704,18 @@ png_create_colormap_entry(png_image_read_control *display,
/* Update the cache with whether the file gamma is significantly different
* from sRGB.
*/
if (encoding == E_FILE)
if (encoding == P_FILE)
{
if (display->file_encoding == E_NOTSET)
if (display->file_encoding == P_NOTSET)
set_file_encoding(display);
/* Note that the cached value may be E_FILE too, but if it is then the
/* Note that the cached value may be P_FILE too, but if it is then the
* gamma_to_linear member has been set.
*/
encoding = display->file_encoding;
}
if (encoding == E_FILE)
if (encoding == P_FILE)
{
png_fixed_point g = display->gamma_to_linear;
@@ -1652,10 +1723,10 @@ png_create_colormap_entry(png_image_read_control *display,
green = png_gamma_16bit_correct(green*257, g);
blue = png_gamma_16bit_correct(blue*257, g);
if (convert_to_Y || output_encoding == E_LINEAR)
if (convert_to_Y || output_encoding == P_LINEAR)
{
alpha *= 257;
encoding = E_LINEAR;
encoding = P_LINEAR;
}
else
@@ -1663,11 +1734,11 @@ png_create_colormap_entry(png_image_read_control *display,
red = PNG_sRGB_FROM_LINEAR(red * 255);
green = PNG_sRGB_FROM_LINEAR(green * 255);
blue = PNG_sRGB_FROM_LINEAR(blue * 255);
encoding = E_sRGB;
encoding = P_sRGB;
}
}
else if (encoding == E_LINEAR8)
else if (encoding == P_LINEAR8)
{
/* This encoding occurs quite frequently in test cases because PngSuite
* includes a gAMA 1.0 chunk with most images.
@@ -1676,10 +1747,10 @@ png_create_colormap_entry(png_image_read_control *display,
green *= 257;
blue *= 257;
alpha *= 257;
encoding = E_LINEAR;
encoding = P_LINEAR;
}
else if (encoding == E_sRGB && (convert_to_Y || output_encoding == E_LINEAR))
else if (encoding == P_sRGB && (convert_to_Y || output_encoding == P_LINEAR))
{
/* The values are 8-bit sRGB values, but must be converted to 16-bit
* linear.
@@ -1688,11 +1759,11 @@ png_create_colormap_entry(png_image_read_control *display,
green = png_sRGB_table[green];
blue = png_sRGB_table[blue];
alpha *= 257;
encoding = E_LINEAR;
encoding = P_LINEAR;
}
/* This is set if the color isn't gray but the output is. */
if (encoding == E_LINEAR)
if (encoding == P_LINEAR)
{
if (convert_to_Y)
{
@@ -1700,7 +1771,7 @@ png_create_colormap_entry(png_image_read_control *display,
png_uint_32 y = (png_uint_32)6968 * red + (png_uint_32)23434 * green +
(png_uint_32)2366 * blue;
if (output_encoding == E_LINEAR)
if (output_encoding == P_LINEAR)
y = (y + 16384) >> 15;
else
@@ -1709,19 +1780,19 @@ png_create_colormap_entry(png_image_read_control *display,
y = (y + 128) >> 8;
y *= 255;
y = PNG_sRGB_FROM_LINEAR((y + 64) >> 7);
encoding = E_sRGB;
encoding = P_sRGB;
}
blue = red = green = y;
}
else if (output_encoding == E_sRGB)
else if (output_encoding == P_sRGB)
{
red = PNG_sRGB_FROM_LINEAR(red * 255);
green = PNG_sRGB_FROM_LINEAR(green * 255);
blue = PNG_sRGB_FROM_LINEAR(blue * 255);
alpha = PNG_DIV257(alpha);
encoding = E_sRGB;
encoding = P_sRGB;
}
}
@@ -1730,7 +1801,7 @@ png_create_colormap_entry(png_image_read_control *display,
/* Store the value. */
{
# ifdef PNG_FORMAT_BGR_SUPPORTED
# ifdef PNG_FORMAT_AFIRST_SUPPORTED
const int afirst = (image->format & PNG_FORMAT_FLAG_AFIRST) != 0 &&
(image->format & PNG_FORMAT_FLAG_ALPHA) != 0;
# else
@@ -1742,7 +1813,7 @@ png_create_colormap_entry(png_image_read_control *display,
# define bgr 0
# endif
if (output_encoding == E_LINEAR)
if (output_encoding == P_LINEAR)
{
png_uint_16p entry = png_voidcast(png_uint_16p, display->colormap);
@@ -1797,7 +1868,7 @@ png_create_colormap_entry(png_image_read_control *display,
}
}
else /* output encoding is E_sRGB */
else /* output encoding is P_sRGB */
{
png_bytep entry = png_voidcast(png_bytep, display->colormap);
@@ -1839,7 +1910,7 @@ make_gray_file_colormap(png_image_read_control *display)
unsigned int i;
for (i=0; i<256; ++i)
png_create_colormap_entry(display, i, i, i, i, 255, E_FILE);
png_create_colormap_entry(display, i, i, i, i, 255, P_FILE);
return i;
}
@@ -1850,7 +1921,7 @@ make_gray_colormap(png_image_read_control *display)
unsigned int i;
for (i=0; i<256; ++i)
png_create_colormap_entry(display, i, i, i, i, 255, E_sRGB);
png_create_colormap_entry(display, i, i, i, i, 255, P_sRGB);
return i;
}
@@ -1889,13 +1960,13 @@ make_ga_colormap(png_image_read_control *display)
while (i < 231)
{
unsigned int gray = (i * 256 + 115) / 231;
png_create_colormap_entry(display, i++, gray, gray, gray, 255, E_sRGB);
png_create_colormap_entry(display, i++, gray, gray, gray, 255, P_sRGB);
}
/* 255 is used here for the component values for consistency with the code
* that undoes premultiplication in pngwrite.c.
*/
png_create_colormap_entry(display, i++, 255, 255, 255, 0, E_sRGB);
png_create_colormap_entry(display, i++, 255, 255, 255, 0, P_sRGB);
for (a=1; a<5; ++a)
{
@@ -1903,7 +1974,7 @@ make_ga_colormap(png_image_read_control *display)
for (g=0; g<6; ++g)
png_create_colormap_entry(display, i++, g*51, g*51, g*51, a*51,
E_sRGB);
P_sRGB);
}
return i;
@@ -1927,7 +1998,7 @@ make_rgb_colormap(png_image_read_control *display)
for (b=0; b<6; ++b)
png_create_colormap_entry(display, i++, r*51, g*51, b*51, 255,
E_sRGB);
P_sRGB);
}
}
@@ -1950,11 +2021,11 @@ png_image_read_colormap(png_voidp argument)
const png_structrp png_ptr = image->opaque->png_ptr;
const png_uint_32 output_format = image->format;
const int output_encoding = (output_format & PNG_FORMAT_FLAG_LINEAR) ?
E_LINEAR : E_sRGB;
P_LINEAR : P_sRGB;
unsigned int cmap_entries;
unsigned int output_processing; /* Output processing option */
unsigned int data_encoding = E_NOTSET; /* Encoding libpng must produce */
unsigned int data_encoding = P_NOTSET; /* Encoding libpng must produce */
/* Background information; the background color and the index of this color
* in the color-map if it exists (else 256).
@@ -1974,7 +2045,7 @@ png_image_read_colormap(png_voidp argument)
png_ptr->num_trans > 0) /* alpha in input */ &&
((output_format & PNG_FORMAT_FLAG_ALPHA) == 0) /* no alpha in output */)
{
if (output_encoding == E_LINEAR) /* compose on black */
if (output_encoding == P_LINEAR) /* compose on black */
back_b = back_g = back_r = 0;
else if (display->background == NULL /* no way to remove it */)
@@ -1998,7 +2069,7 @@ png_image_read_colormap(png_voidp argument)
}
}
else if (output_encoding == E_LINEAR)
else if (output_encoding == P_LINEAR)
back_b = back_r = back_g = 65535;
else
@@ -2056,7 +2127,7 @@ png_image_read_colormap(png_voidp argument)
trans = png_ptr->trans_color.gray;
if ((output_format & PNG_FORMAT_FLAG_ALPHA) == 0)
back_alpha = output_encoding == E_LINEAR ? 65535 : 255;
back_alpha = output_encoding == P_LINEAR ? 65535 : 255;
}
/* png_create_colormap_entry just takes an RGBA and writes the
@@ -2074,7 +2145,7 @@ png_image_read_colormap(png_voidp argument)
*/
if (i != trans)
png_create_colormap_entry(display, i, val, val, val, 255,
E_FILE/*8-bit with file gamma*/);
P_FILE/*8-bit with file gamma*/);
/* Else this entry is transparent. The colors don't matter if
* there is an alpha channel (back_alpha == 0), but it does no
@@ -2090,7 +2161,7 @@ png_image_read_colormap(png_voidp argument)
}
/* We need libpng to preserve the original encoding. */
data_encoding = E_FILE;
data_encoding = P_FILE;
/* The rows from libpng, while technically gray values, are now also
* color-map indicies; however, they may need to be expanded to 1
@@ -2119,7 +2190,7 @@ png_image_read_colormap(png_voidp argument)
* ensuring that the corresponding gray level matches the background
* color exactly.
*/
data_encoding = E_sRGB;
data_encoding = P_sRGB;
if (PNG_GRAY_COLORMAP_ENTRIES > image->colormap_entries)
png_error(png_ptr, "gray[16] color-map: too few entries");
@@ -2143,7 +2214,7 @@ png_image_read_colormap(png_voidp argument)
png_color_16 c;
png_uint_32 gray = back_g;
if (output_encoding == E_LINEAR)
if (output_encoding == P_LINEAR)
{
gray = PNG_sRGB_FROM_LINEAR(gray * 255);
@@ -2151,7 +2222,7 @@ png_image_read_colormap(png_voidp argument)
* matches.
*/
png_create_colormap_entry(display, gray, back_g, back_g,
back_g, 65535, E_LINEAR);
back_g, 65535, P_LINEAR);
}
/* The background passed to libpng, however, must be the
@@ -2172,7 +2243,7 @@ png_image_read_colormap(png_voidp argument)
break;
}
back_alpha = output_encoding == E_LINEAR ? 65535 : 255;
back_alpha = output_encoding == P_LINEAR ? 65535 : 255;
}
/* output_processing means that the libpng-processed row will be
@@ -2209,7 +2280,7 @@ png_image_read_colormap(png_voidp argument)
* worry about tRNS matching - tRNS is ignored if there is an alpha
* channel.
*/
data_encoding = E_sRGB;
data_encoding = P_sRGB;
if (output_format & PNG_FORMAT_FLAG_ALPHA)
{
@@ -2252,13 +2323,13 @@ png_image_read_colormap(png_voidp argument)
cmap_entries = make_gray_colormap(display);
if (output_encoding == E_LINEAR)
if (output_encoding == P_LINEAR)
{
gray = PNG_sRGB_FROM_LINEAR(gray * 255);
/* And make sure the corresponding palette entry matches. */
png_create_colormap_entry(display, gray, back_g, back_g,
back_g, 65535, E_LINEAR);
back_g, 65535, P_LINEAR);
}
/* The background passed to libpng, however, must be the sRGB
@@ -2289,7 +2360,7 @@ png_image_read_colormap(png_voidp argument)
{
png_uint_32 gray = (i * 256 + 115) / 231;
png_create_colormap_entry(display, i++, gray, gray, gray,
255, E_sRGB);
255, P_sRGB);
}
/* NOTE: this preserves the full precision of the application
@@ -2297,7 +2368,7 @@ png_image_read_colormap(png_voidp argument)
*/
background_index = i;
png_create_colormap_entry(display, i++, back_r, back_g, back_b,
output_encoding == E_LINEAR ? 65535U : 255U, output_encoding);
output_encoding == P_LINEAR ? 65535U : 255U, output_encoding);
/* For non-opaque input composite on the sRGB background - this
* requires inverting the encoding for each component. The input
@@ -2307,7 +2378,7 @@ png_image_read_colormap(png_voidp argument)
* represents. Consequently 'G' is always sRGB encoded, while
* 'A' is linear. We need the linear background colors.
*/
if (output_encoding == E_sRGB) /* else already linear */
if (output_encoding == P_sRGB) /* else already linear */
{
/* This may produce a value not exactly matching the
* background, but that's ok because these numbers are only
@@ -2337,7 +2408,7 @@ png_image_read_colormap(png_voidp argument)
png_create_colormap_entry(display, i++,
PNG_sRGB_FROM_LINEAR(gray + back_rx),
PNG_sRGB_FROM_LINEAR(gray + back_gx),
PNG_sRGB_FROM_LINEAR(gray + back_bx), 255, E_sRGB);
PNG_sRGB_FROM_LINEAR(gray + back_bx), 255, P_sRGB);
}
}
@@ -2364,7 +2435,7 @@ png_image_read_colormap(png_voidp argument)
*/
png_set_rgb_to_gray_fixed(png_ptr, PNG_ERROR_ACTION_NONE, -1,
-1);
data_encoding = E_sRGB;
data_encoding = P_sRGB;
/* The output will now be one or two 8-bit gray or gray+alpha
* channels. The more complex case arises when the input has alpha.
@@ -2409,7 +2480,7 @@ png_image_read_colormap(png_voidp argument)
png_gamma_not_sRGB(png_ptr->colorspace.gamma))
{
cmap_entries = make_gray_file_colormap(display);
data_encoding = E_FILE;
data_encoding = P_FILE;
}
else
@@ -2428,18 +2499,18 @@ png_image_read_colormap(png_voidp argument)
* it. Achieve this simply by ensuring that the entry
* selected for the background really is the background color.
*/
if (data_encoding == E_FILE) /* from the fixup above */
if (data_encoding == P_FILE) /* from the fixup above */
{
/* The app supplied a gray which is in output_encoding, we
* need to convert it to a value of the input (E_FILE)
* need to convert it to a value of the input (P_FILE)
* encoding then set this palette entry to the required
* output encoding.
*/
if (output_encoding == E_sRGB)
gray = png_sRGB_table[gray]; /* now E_LINEAR */
if (output_encoding == P_sRGB)
gray = png_sRGB_table[gray]; /* now P_LINEAR */
gray = PNG_DIV257(png_gamma_16bit_correct(gray,
png_ptr->colorspace.gamma)); /* now E_FILE */
png_ptr->colorspace.gamma)); /* now P_FILE */
/* And make sure the corresponding palette entry contains
* exactly the required sRGB value.
@@ -2448,14 +2519,14 @@ png_image_read_colormap(png_voidp argument)
back_g, 0/*unused*/, output_encoding);
}
else if (output_encoding == E_LINEAR)
else if (output_encoding == P_LINEAR)
{
gray = PNG_sRGB_FROM_LINEAR(gray * 255);
/* And make sure the corresponding palette entry matches.
*/
png_create_colormap_entry(display, gray, back_g, back_g,
back_g, 0/*unused*/, E_LINEAR);
back_g, 0/*unused*/, P_LINEAR);
}
/* The background passed to libpng, however, must be the
@@ -2485,7 +2556,7 @@ png_image_read_colormap(png_voidp argument)
* to do it once and using PNG_DIV51 on the 6x6x6 reduced RGB cube.
* Consequently we always want libpng to produce sRGB data.
*/
data_encoding = E_sRGB;
data_encoding = P_sRGB;
/* Is there any transparency or alpha? */
if (png_ptr->color_type == PNG_COLOR_TYPE_RGB_ALPHA ||
@@ -2505,7 +2576,7 @@ png_image_read_colormap(png_voidp argument)
/* Add a transparent entry. */
png_create_colormap_entry(display, cmap_entries, 255, 255,
255, 0, E_sRGB);
255, 0, P_sRGB);
/* This is stored as the background index for the processing
* algorithm.
@@ -2526,7 +2597,7 @@ png_image_read_colormap(png_voidp argument)
*/
for (b=0; b<256; b = (b << 1) | 0x7f)
png_create_colormap_entry(display, cmap_entries++,
r, g, b, 128, E_sRGB);
r, g, b, 128, P_sRGB);
}
}
@@ -2555,7 +2626,7 @@ png_image_read_colormap(png_voidp argument)
png_create_colormap_entry(display, cmap_entries, back_r,
back_g, back_b, 0/*unused*/, output_encoding);
if (output_encoding == E_LINEAR)
if (output_encoding == P_LINEAR)
{
r = PNG_sRGB_FROM_LINEAR(back_r * 255);
g = PNG_sRGB_FROM_LINEAR(back_g * 255);
@@ -2595,11 +2666,11 @@ png_image_read_colormap(png_voidp argument)
*/
for (b=0; b<256; b = (b << 1) | 0x7f)
png_create_colormap_entry(display, cmap_entries++,
png_colormap_compose(display, r, E_sRGB, 128,
png_colormap_compose(display, r, P_sRGB, 128,
back_r, output_encoding),
png_colormap_compose(display, g, E_sRGB, 128,
png_colormap_compose(display, g, P_sRGB, 128,
back_g, output_encoding),
png_colormap_compose(display, b, E_sRGB, 128,
png_colormap_compose(display, b, P_sRGB, 128,
back_b, output_encoding),
0/*unused*/, output_encoding);
}
@@ -2658,7 +2729,7 @@ png_image_read_colormap(png_voidp argument)
num_trans = 0;
output_processing = PNG_CMAP_NONE;
data_encoding = E_FILE; /* Don't change from color-map indicies */
data_encoding = P_FILE; /* Don't change from color-map indicies */
cmap_entries = png_ptr->num_palette;
if (cmap_entries > 256)
cmap_entries = 256;
@@ -2680,13 +2751,13 @@ png_image_read_colormap(png_voidp argument)
* on the sRGB color in 'back'.
*/
png_create_colormap_entry(display, i,
png_colormap_compose(display, colormap[i].red, E_FILE,
png_colormap_compose(display, colormap[i].red, P_FILE,
trans[i], back_r, output_encoding),
png_colormap_compose(display, colormap[i].green, E_FILE,
png_colormap_compose(display, colormap[i].green, P_FILE,
trans[i], back_g, output_encoding),
png_colormap_compose(display, colormap[i].blue, E_FILE,
png_colormap_compose(display, colormap[i].blue, P_FILE,
trans[i], back_b, output_encoding),
output_encoding == E_LINEAR ? trans[i] * 257U :
output_encoding == P_LINEAR ? trans[i] * 257U :
trans[i],
output_encoding);
}
@@ -2695,7 +2766,7 @@ png_image_read_colormap(png_voidp argument)
else
png_create_colormap_entry(display, i, colormap[i].red,
colormap[i].green, colormap[i].blue,
i < num_trans ? trans[i] : 255U, E_FILE/*8-bit*/);
i < num_trans ? trans[i] : 255U, P_FILE/*8-bit*/);
}
/* The PNG data may have indicies packed in fewer than 8 bits, it
@@ -2723,12 +2794,12 @@ png_image_read_colormap(png_voidp argument)
png_error(png_ptr, "bad data option (internal error)");
break;
case E_sRGB:
case P_sRGB:
/* Change to 8-bit sRGB */
png_set_alpha_mode_fixed(png_ptr, PNG_ALPHA_PNG, PNG_GAMMA_sRGB);
/* FALL THROUGH */
case E_FILE:
case P_FILE:
if (png_ptr->bit_depth > 8)
png_set_scale_16(png_ptr);
break;
@@ -2805,7 +2876,6 @@ png_image_read_and_map(png_voidp argument)
break;
default:
passes = 0;
png_error(png_ptr, "unknown interlace type");
}
@@ -3124,7 +3194,6 @@ png_image_read_composite(png_voidp argument)
break;
default:
passes = 0;
png_error(png_ptr, "unknown interlace type");
}
@@ -3273,11 +3342,15 @@ png_image_read_background(png_voidp argument)
break;
default:
passes = 0;
png_error(png_ptr, "unknown interlace type");
}
switch (png_get_bit_depth(png_ptr, info_ptr))
/* Use direct access to info_ptr here because otherwise the simplified API
* would require PNG_EASY_ACCESS_SUPPORTED (just for this.) Note this is
* checking the value after libpng expansions, not the original value in the
* PNG.
*/
switch (info_ptr->bit_depth)
{
default:
png_error(png_ptr, "unexpected bit depth");
@@ -3425,8 +3498,10 @@ png_image_read_background(png_voidp argument)
unsigned int outchannels = 1+preserve_alpha;
int swap_alpha = 0;
if (preserve_alpha && (image->format & PNG_FORMAT_FLAG_AFIRST))
swap_alpha = 1;
# ifdef PNG_SIMPLIFIED_READ_AFIRST_SUPPORTED
if (preserve_alpha && (image->format & PNG_FORMAT_FLAG_AFIRST))
swap_alpha = 1;
# endif
for (pass = 0; pass < passes; ++pass)
{

View File

@@ -1839,6 +1839,9 @@ png_read_transform_info(png_structrp png_ptr, png_inforp info_ptr)
info_ptr->bit_depth = 8;
info_ptr->num_trans = 0;
if (png_ptr->palette == NULL)
png_error (png_ptr, "Palette is NULL in indexed image");
}
else
{
@@ -2017,303 +2020,6 @@ defined(PNG_READ_USER_TRANSFORM_SUPPORTED)
#endif
}
/* Transform the row. The order of transformations is significant,
* and is very touchy. If you add a transformation, take care to
* decide how it fits in with the other transformations here.
*/
void /* PRIVATE */
png_do_read_transformations(png_structrp png_ptr, png_row_infop row_info)
{
png_debug(1, "in png_do_read_transformations");
if (png_ptr->row_buf == NULL)
{
/* Prior to 1.5.4 this output row/pass where the NULL pointer is, but this
* error is incredibly rare and incredibly easy to debug without this
* information.
*/
png_error(png_ptr, "NULL row buffer");
}
/* The following is debugging; prior to 1.5.4 the code was never compiled in;
* in 1.5.4 PNG_FLAG_DETECT_UNINITIALIZED was added and the macro
* PNG_WARN_UNINITIALIZED_ROW removed. In 1.6 the new flag is set only for
* all transformations, however in practice the ROW_INIT always gets done on
* demand, if necessary.
*/
if ((png_ptr->flags & PNG_FLAG_DETECT_UNINITIALIZED) != 0 &&
!(png_ptr->flags & PNG_FLAG_ROW_INIT))
{
/* Application has failed to call either png_read_start_image() or
* png_read_update_info() after setting transforms that expand pixels.
* This check added to libpng-1.2.19 (but not enabled until 1.5.4).
*/
png_error(png_ptr, "Uninitialized row");
}
#ifdef PNG_READ_EXPAND_SUPPORTED
if (png_ptr->transformations & PNG_EXPAND)
{
if (row_info->color_type == PNG_COLOR_TYPE_PALETTE)
{
png_do_expand_palette(row_info, png_ptr->row_buf + 1,
png_ptr->palette, png_ptr->trans_alpha, png_ptr->num_trans);
}
else
{
if (png_ptr->num_trans &&
(png_ptr->transformations & PNG_EXPAND_tRNS))
png_do_expand(row_info, png_ptr->row_buf + 1,
&(png_ptr->trans_color));
else
png_do_expand(row_info, png_ptr->row_buf + 1,
NULL);
}
}
#endif
#ifdef PNG_READ_STRIP_ALPHA_SUPPORTED
if ((png_ptr->transformations & PNG_STRIP_ALPHA) &&
!(png_ptr->transformations & PNG_COMPOSE) &&
(row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA ||
row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA))
png_do_strip_channel(row_info, png_ptr->row_buf + 1,
0 /* at_start == false, because SWAP_ALPHA happens later */);
#endif
#ifdef PNG_READ_RGB_TO_GRAY_SUPPORTED
if (png_ptr->transformations & PNG_RGB_TO_GRAY)
{
int rgb_error =
png_do_rgb_to_gray(png_ptr, row_info,
png_ptr->row_buf + 1);
if (rgb_error)
{
png_ptr->rgb_to_gray_status=1;
if ((png_ptr->transformations & PNG_RGB_TO_GRAY) ==
PNG_RGB_TO_GRAY_WARN)
png_warning(png_ptr, "png_do_rgb_to_gray found nongray pixel");
if ((png_ptr->transformations & PNG_RGB_TO_GRAY) ==
PNG_RGB_TO_GRAY_ERR)
png_error(png_ptr, "png_do_rgb_to_gray found nongray pixel");
}
}
#endif
/* From Andreas Dilger e-mail to png-implement, 26 March 1998:
*
* In most cases, the "simple transparency" should be done prior to doing
* gray-to-RGB, or you will have to test 3x as many bytes to check if a
* pixel is transparent. You would also need to make sure that the
* transparency information is upgraded to RGB.
*
* To summarize, the current flow is:
* - Gray + simple transparency -> compare 1 or 2 gray bytes and composite
* with background "in place" if transparent,
* convert to RGB if necessary
* - Gray + alpha -> composite with gray background and remove alpha bytes,
* convert to RGB if necessary
*
* To support RGB backgrounds for gray images we need:
* - Gray + simple transparency -> convert to RGB + simple transparency,
* compare 3 or 6 bytes and composite with
* background "in place" if transparent
* (3x compare/pixel compared to doing
* composite with gray bkgrnd)
* - Gray + alpha -> convert to RGB + alpha, composite with background and
* remove alpha bytes (3x float
* operations/pixel compared with composite
* on gray background)
*
* Greg's change will do this. The reason it wasn't done before is for
* performance, as this increases the per-pixel operations. If we would check
* in advance if the background was gray or RGB, and position the gray-to-RGB
* transform appropriately, then it would save a lot of work/time.
*/
#ifdef PNG_READ_GRAY_TO_RGB_SUPPORTED
/* If gray -> RGB, do so now only if background is non-gray; else do later
* for performance reasons
*/
if ((png_ptr->transformations & PNG_GRAY_TO_RGB) &&
!(png_ptr->flags & PNG_FLAG_BACKGROUND_IS_GRAY))
png_do_gray_to_rgb(row_info, png_ptr->row_buf + 1);
#endif
#if defined(PNG_READ_BACKGROUND_SUPPORTED) || \
defined(PNG_READ_ALPHA_MODE_SUPPORTED)
if (png_ptr->transformations & PNG_COMPOSE)
png_do_compose(row_info, png_ptr->row_buf + 1, png_ptr);
#endif
#ifdef PNG_READ_GAMMA_SUPPORTED
if ((png_ptr->transformations & PNG_GAMMA) &&
#ifdef PNG_READ_RGB_TO_GRAY_SUPPORTED
/* Because RGB_TO_GRAY does the gamma transform. */
!(png_ptr->transformations & PNG_RGB_TO_GRAY) &&
#endif
#if defined(PNG_READ_BACKGROUND_SUPPORTED) ||\
defined(PNG_READ_ALPHA_MODE_SUPPORTED)
/* Because PNG_COMPOSE does the gamma transform if there is something to
* do (if there is an alpha channel or transparency.)
*/
!((png_ptr->transformations & PNG_COMPOSE) &&
((png_ptr->num_trans != 0) ||
(png_ptr->color_type & PNG_COLOR_MASK_ALPHA))) &&
#endif
/* Because png_init_read_transformations transforms the palette, unless
* RGB_TO_GRAY will do the transform.
*/
(png_ptr->color_type != PNG_COLOR_TYPE_PALETTE))
png_do_gamma(row_info, png_ptr->row_buf + 1, png_ptr);
#endif
#ifdef PNG_READ_STRIP_ALPHA_SUPPORTED
if ((png_ptr->transformations & PNG_STRIP_ALPHA) &&
(png_ptr->transformations & PNG_COMPOSE) &&
(row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA ||
row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA))
png_do_strip_channel(row_info, png_ptr->row_buf + 1,
0 /* at_start == false, because SWAP_ALPHA happens later */);
#endif
#ifdef PNG_READ_ALPHA_MODE_SUPPORTED
if ((png_ptr->transformations & PNG_ENCODE_ALPHA) &&
(row_info->color_type & PNG_COLOR_MASK_ALPHA))
png_do_encode_alpha(row_info, png_ptr->row_buf + 1, png_ptr);
#endif
#ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
if (png_ptr->transformations & PNG_SCALE_16_TO_8)
png_do_scale_16_to_8(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_STRIP_16_TO_8_SUPPORTED
/* There is no harm in doing both of these because only one has any effect,
* by putting the 'scale' option first if the app asks for scale (either by
* calling the API or in a TRANSFORM flag) this is what happens.
*/
if (png_ptr->transformations & PNG_16_TO_8)
png_do_chop(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_QUANTIZE_SUPPORTED
if (png_ptr->transformations & PNG_QUANTIZE)
{
png_do_quantize(row_info, png_ptr->row_buf + 1,
png_ptr->palette_lookup, png_ptr->quantize_index);
if (row_info->rowbytes == 0)
png_error(png_ptr, "png_do_quantize returned rowbytes=0");
}
#endif /* PNG_READ_QUANTIZE_SUPPORTED */
#ifdef PNG_READ_EXPAND_16_SUPPORTED
/* Do the expansion now, after all the arithmetic has been done. Notice
* that previous transformations can handle the PNG_EXPAND_16 flag if this
* is efficient (particularly true in the case of gamma correction, where
* better accuracy results faster!)
*/
if (png_ptr->transformations & PNG_EXPAND_16)
png_do_expand_16(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_GRAY_TO_RGB_SUPPORTED
/* NOTE: moved here in 1.5.4 (from much later in this list.) */
if ((png_ptr->transformations & PNG_GRAY_TO_RGB) &&
(png_ptr->flags & PNG_FLAG_BACKGROUND_IS_GRAY))
png_do_gray_to_rgb(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_INVERT_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_MONO)
png_do_invert(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_SHIFT_SUPPORTED
if (png_ptr->transformations & PNG_SHIFT)
png_do_unshift(row_info, png_ptr->row_buf + 1,
&(png_ptr->shift));
#endif
#ifdef PNG_READ_PACK_SUPPORTED
if (png_ptr->transformations & PNG_PACK)
png_do_unpack(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_CHECK_FOR_INVALID_INDEX_SUPPORTED
/* Added at libpng-1.5.10 */
if (row_info->color_type == PNG_COLOR_TYPE_PALETTE &&
png_ptr->num_palette_max >= 0)
png_do_check_palette_indexes(png_ptr, row_info);
#endif
#ifdef PNG_READ_BGR_SUPPORTED
if (png_ptr->transformations & PNG_BGR)
png_do_bgr(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_PACKSWAP_SUPPORTED
if (png_ptr->transformations & PNG_PACKSWAP)
png_do_packswap(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_FILLER_SUPPORTED
if (png_ptr->transformations & PNG_FILLER)
png_do_read_filler(row_info, png_ptr->row_buf + 1,
(png_uint_32)png_ptr->filler, png_ptr->flags);
#endif
#ifdef PNG_READ_INVERT_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_ALPHA)
png_do_read_invert_alpha(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_SWAP_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_ALPHA)
png_do_read_swap_alpha(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_16BIT_SUPPORTED
#ifdef PNG_READ_SWAP_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_BYTES)
png_do_swap(row_info, png_ptr->row_buf + 1);
#endif
#endif
#ifdef PNG_READ_USER_TRANSFORM_SUPPORTED
if (png_ptr->transformations & PNG_USER_TRANSFORM)
{
if (png_ptr->read_user_transform_fn != NULL)
(*(png_ptr->read_user_transform_fn)) /* User read transform function */
(png_ptr, /* png_ptr */
row_info, /* row_info: */
/* png_uint_32 width; width of row */
/* png_size_t rowbytes; number of bytes in row */
/* png_byte color_type; color type of pixels */
/* png_byte bit_depth; bit depth of samples */
/* png_byte channels; number of channels (1-4) */
/* png_byte pixel_depth; bits per pixel (depth*channels) */
png_ptr->row_buf + 1); /* start of pixel data for row */
#ifdef PNG_USER_TRANSFORM_PTR_SUPPORTED
if (png_ptr->user_transform_depth)
row_info->bit_depth = png_ptr->user_transform_depth;
if (png_ptr->user_transform_channels)
row_info->channels = png_ptr->user_transform_channels;
#endif
row_info->pixel_depth = (png_byte)(row_info->bit_depth *
row_info->channels);
row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth, row_info->width);
}
#endif
}
#ifdef PNG_READ_PACK_SUPPORTED
/* Unpack pixels of 1, 2, or 4 bits per pixel into 1 byte per pixel,
* without changing the actual values. Thus, if you had a row with
@@ -2321,7 +2027,7 @@ png_do_read_transformations(png_structrp png_ptr, png_row_infop row_info)
* the numbers 0 or 1. If you would rather they contain 0 and 255, use
* png_do_shift() after this.
*/
void /* PRIVATE */
static void
png_do_unpack(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_unpack");
@@ -2419,7 +2125,7 @@ png_do_unpack(png_row_infop row_info, png_bytep row)
* a row of bit depth 8, but only 5 are significant, this will shift
* the values back to 0 through 31.
*/
void /* PRIVATE */
static void
png_do_unshift(png_row_infop row_info, png_bytep row,
png_const_color_8p sig_bits)
{
@@ -2558,7 +2264,7 @@ png_do_unshift(png_row_infop row_info, png_bytep row,
#ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
/* Scale rows of bit depth 16 down to 8 accurately */
void /* PRIVATE */
static void
png_do_scale_16_to_8(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_scale_16_to_8");
@@ -2616,7 +2322,7 @@ png_do_scale_16_to_8(png_row_infop row_info, png_bytep row)
#endif
#ifdef PNG_READ_STRIP_16_TO_8_SUPPORTED
void /* PRIVATE */
static void
/* Simply discard the low byte. This was the default behavior prior
* to libpng-1.5.4.
*/
@@ -2644,7 +2350,7 @@ png_do_chop(png_row_infop row_info, png_bytep row)
#endif
#ifdef PNG_READ_SWAP_ALPHA_SUPPORTED
void /* PRIVATE */
static void
png_do_read_swap_alpha(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_read_swap_alpha");
@@ -2741,7 +2447,7 @@ png_do_read_swap_alpha(png_row_infop row_info, png_bytep row)
#endif
#ifdef PNG_READ_INVERT_ALPHA_SUPPORTED
void /* PRIVATE */
static void
png_do_read_invert_alpha(png_row_infop row_info, png_bytep row)
{
png_uint_32 row_width;
@@ -2843,7 +2549,7 @@ png_do_read_invert_alpha(png_row_infop row_info, png_bytep row)
#ifdef PNG_READ_FILLER_SUPPORTED
/* Add filler channel if we have RGB color */
void /* PRIVATE */
static void
png_do_read_filler(png_row_infop row_info, png_bytep row,
png_uint_32 filler, png_uint_32 flags)
{
@@ -3030,7 +2736,7 @@ png_do_read_filler(png_row_infop row_info, png_bytep row,
#ifdef PNG_READ_GRAY_TO_RGB_SUPPORTED
/* Expand grayscale files to RGB, with or without alpha */
void /* PRIVATE */
static void
png_do_gray_to_rgb(png_row_infop row_info, png_bytep row)
{
png_uint_32 i;
@@ -3169,7 +2875,7 @@ png_do_gray_to_rgb(png_row_infop row_info, png_bytep row)
* calculated to make the sum 32768. This will result in different rounding
* to that used above.
*/
int /* PRIVATE */
static int
png_do_rgb_to_gray(png_structrp png_ptr, png_row_infop row_info, png_bytep row)
{
@@ -3358,73 +3064,14 @@ png_do_rgb_to_gray(png_structrp png_ptr, png_row_infop row_info, png_bytep row)
return rgb_error;
}
#endif
#endif /* PNG_READ_TRANSFORMS_SUPPORTED */
#ifdef PNG_BUILD_GRAYSCALE_PALETTE_SUPPORTED
/* Build a grayscale palette. Palette is assumed to be 1 << bit_depth
* large of png_color. This lets grayscale images be treated as
* paletted. Most useful for gamma correction and simplification
* of code. This API is not used internally.
*/
void PNGAPI
png_build_grayscale_palette(int bit_depth, png_colorp palette)
{
int num_palette;
int color_inc;
int i;
int v;
png_debug(1, "in png_do_build_grayscale_palette");
if (palette == NULL)
return;
switch (bit_depth)
{
case 1:
num_palette = 2;
color_inc = 0xff;
break;
case 2:
num_palette = 4;
color_inc = 0x55;
break;
case 4:
num_palette = 16;
color_inc = 0x11;
break;
case 8:
num_palette = 256;
color_inc = 1;
break;
default:
num_palette = 0;
color_inc = 0;
break;
}
for (i = 0, v = 0; i < num_palette; i++, v += color_inc)
{
palette[i].red = (png_byte)v;
palette[i].green = (png_byte)v;
palette[i].blue = (png_byte)v;
}
}
#endif
#ifdef PNG_READ_TRANSFORMS_SUPPORTED
#if defined(PNG_READ_BACKGROUND_SUPPORTED) ||\
defined(PNG_READ_ALPHA_MODE_SUPPORTED)
/* Replace any alpha or transparency with the supplied background color.
* "background" is already in the screen gamma, while "background_1" is
* at a gamma of 1.0. Paletted files have already been taken care of.
*/
void /* PRIVATE */
static void
png_do_compose(png_row_infop row_info, png_bytep row, png_structrp png_ptr)
{
png_const_bytep gamma_table = png_ptr->gamma_table;
@@ -4186,7 +3833,7 @@ png_do_compose(png_row_infop row_info, png_bytep row, png_structrp png_ptr)
* is 16, use gamma_16_table and gamma_shift. Build these with
* build_gamma_table().
*/
void /* PRIVATE */
static void
png_do_gamma(png_row_infop row_info, png_bytep row, png_structrp png_ptr)
{
png_const_bytep gamma_table = png_ptr->gamma_table;
@@ -4366,7 +4013,7 @@ png_do_gamma(png_row_infop row_info, png_bytep row, png_structrp png_ptr)
* linear.) Called only with color types that have an alpha channel. Needs the
* from_1 tables.
*/
void /* PRIVATE */
static void
png_do_encode_alpha(png_row_infop row_info, png_bytep row, png_structrp png_ptr)
{
png_uint_32 row_width = row_info->width;
@@ -4434,7 +4081,7 @@ png_do_encode_alpha(png_row_infop row_info, png_bytep row, png_structrp png_ptr)
/* Expands a palette row to an RGB or RGBA row depending
* upon whether you supply trans and num_trans.
*/
void /* PRIVATE */
static void
png_do_expand_palette(png_row_infop row_info, png_bytep row,
png_const_colorp palette, png_const_bytep trans_alpha, int num_trans)
{
@@ -4587,7 +4234,7 @@ png_do_expand_palette(png_row_infop row_info, png_bytep row,
/* If the bit depth < 8, it is expanded to 8. Also, if the already
* expanded transparency value is supplied, an alpha channel is built.
*/
void /* PRIVATE */
static void
png_do_expand(png_row_infop row_info, png_bytep row,
png_const_color_16p trans_color)
{
@@ -4817,7 +4464,7 @@ png_do_expand(png_row_infop row_info, png_bytep row,
/* If the bit depth is 8 and the color type is not a palette type expand the
* whole row to 16 bits. Has no effect otherwise.
*/
void /* PRIVATE */
static void
png_do_expand_16(png_row_infop row_info, png_bytep row)
{
if (row_info->bit_depth == 8 &&
@@ -4845,7 +4492,7 @@ png_do_expand_16(png_row_infop row_info, png_bytep row)
#endif
#ifdef PNG_READ_QUANTIZE_SUPPORTED
void /* PRIVATE */
static void
png_do_quantize(png_row_infop row_info, png_bytep row,
png_const_bytep palette_lookup, png_const_bytep quantize_lookup)
{
@@ -4937,69 +4584,302 @@ png_do_quantize(png_row_infop row_info, png_bytep row,
}
}
#endif /* PNG_READ_QUANTIZE_SUPPORTED */
#endif /* PNG_READ_TRANSFORMS_SUPPORTED */
#ifdef PNG_MNG_FEATURES_SUPPORTED
/* Undoes intrapixel differencing */
/* Transform the row. The order of transformations is significant,
* and is very touchy. If you add a transformation, take care to
* decide how it fits in with the other transformations here.
*/
void /* PRIVATE */
png_do_read_intrapixel(png_row_infop row_info, png_bytep row)
png_do_read_transformations(png_structrp png_ptr, png_row_infop row_info)
{
png_debug(1, "in png_do_read_intrapixel");
png_debug(1, "in png_do_read_transformations");
if (
(row_info->color_type & PNG_COLOR_MASK_COLOR))
if (png_ptr->row_buf == NULL)
{
int bytes_per_pixel;
png_uint_32 row_width = row_info->width;
/* Prior to 1.5.4 this output row/pass where the NULL pointer is, but this
* error is incredibly rare and incredibly easy to debug without this
* information.
*/
png_error(png_ptr, "NULL row buffer");
}
if (row_info->bit_depth == 8)
/* The following is debugging; prior to 1.5.4 the code was never compiled in;
* in 1.5.4 PNG_FLAG_DETECT_UNINITIALIZED was added and the macro
* PNG_WARN_UNINITIALIZED_ROW removed. In 1.6 the new flag is set only for
* all transformations, however in practice the ROW_INIT always gets done on
* demand, if necessary.
*/
if ((png_ptr->flags & PNG_FLAG_DETECT_UNINITIALIZED) != 0 &&
!(png_ptr->flags & PNG_FLAG_ROW_INIT))
{
/* Application has failed to call either png_read_start_image() or
* png_read_update_info() after setting transforms that expand pixels.
* This check added to libpng-1.2.19 (but not enabled until 1.5.4).
*/
png_error(png_ptr, "Uninitialized row");
}
#ifdef PNG_READ_EXPAND_SUPPORTED
if (png_ptr->transformations & PNG_EXPAND)
{
if (row_info->color_type == PNG_COLOR_TYPE_PALETTE)
{
png_bytep rp;
png_uint_32 i;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 3;
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 4;
else
return;
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
*(rp) = (png_byte)((256 + *rp + *(rp + 1)) & 0xff);
*(rp+2) = (png_byte)((256 + *(rp + 2) + *(rp + 1)) & 0xff);
}
png_do_expand_palette(row_info, png_ptr->row_buf + 1,
png_ptr->palette, png_ptr->trans_alpha, png_ptr->num_trans);
}
else if (row_info->bit_depth == 16)
else
{
png_bytep rp;
png_uint_32 i;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 6;
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 8;
if (png_ptr->num_trans &&
(png_ptr->transformations & PNG_EXPAND_tRNS))
png_do_expand(row_info, png_ptr->row_buf + 1,
&(png_ptr->trans_color));
else
return;
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
png_uint_32 s0 = (*(rp ) << 8) | *(rp + 1);
png_uint_32 s1 = (*(rp + 2) << 8) | *(rp + 3);
png_uint_32 s2 = (*(rp + 4) << 8) | *(rp + 5);
png_uint_32 red = (s0 + s1 + 65536) & 0xffff;
png_uint_32 blue = (s2 + s1 + 65536) & 0xffff;
*(rp ) = (png_byte)((red >> 8) & 0xff);
*(rp + 1) = (png_byte)(red & 0xff);
*(rp + 4) = (png_byte)((blue >> 8) & 0xff);
*(rp + 5) = (png_byte)(blue & 0xff);
}
png_do_expand(row_info, png_ptr->row_buf + 1,
NULL);
}
}
#endif
#ifdef PNG_READ_STRIP_ALPHA_SUPPORTED
if ((png_ptr->transformations & PNG_STRIP_ALPHA) &&
!(png_ptr->transformations & PNG_COMPOSE) &&
(row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA ||
row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA))
png_do_strip_channel(row_info, png_ptr->row_buf + 1,
0 /* at_start == false, because SWAP_ALPHA happens later */);
#endif
#ifdef PNG_READ_RGB_TO_GRAY_SUPPORTED
if (png_ptr->transformations & PNG_RGB_TO_GRAY)
{
int rgb_error =
png_do_rgb_to_gray(png_ptr, row_info,
png_ptr->row_buf + 1);
if (rgb_error)
{
png_ptr->rgb_to_gray_status=1;
if ((png_ptr->transformations & PNG_RGB_TO_GRAY) ==
PNG_RGB_TO_GRAY_WARN)
png_warning(png_ptr, "png_do_rgb_to_gray found nongray pixel");
if ((png_ptr->transformations & PNG_RGB_TO_GRAY) ==
PNG_RGB_TO_GRAY_ERR)
png_error(png_ptr, "png_do_rgb_to_gray found nongray pixel");
}
}
#endif
/* From Andreas Dilger e-mail to png-implement, 26 March 1998:
*
* In most cases, the "simple transparency" should be done prior to doing
* gray-to-RGB, or you will have to test 3x as many bytes to check if a
* pixel is transparent. You would also need to make sure that the
* transparency information is upgraded to RGB.
*
* To summarize, the current flow is:
* - Gray + simple transparency -> compare 1 or 2 gray bytes and composite
* with background "in place" if transparent,
* convert to RGB if necessary
* - Gray + alpha -> composite with gray background and remove alpha bytes,
* convert to RGB if necessary
*
* To support RGB backgrounds for gray images we need:
* - Gray + simple transparency -> convert to RGB + simple transparency,
* compare 3 or 6 bytes and composite with
* background "in place" if transparent
* (3x compare/pixel compared to doing
* composite with gray bkgrnd)
* - Gray + alpha -> convert to RGB + alpha, composite with background and
* remove alpha bytes (3x float
* operations/pixel compared with composite
* on gray background)
*
* Greg's change will do this. The reason it wasn't done before is for
* performance, as this increases the per-pixel operations. If we would check
* in advance if the background was gray or RGB, and position the gray-to-RGB
* transform appropriately, then it would save a lot of work/time.
*/
#ifdef PNG_READ_GRAY_TO_RGB_SUPPORTED
/* If gray -> RGB, do so now only if background is non-gray; else do later
* for performance reasons
*/
if ((png_ptr->transformations & PNG_GRAY_TO_RGB) &&
!(png_ptr->flags & PNG_FLAG_BACKGROUND_IS_GRAY))
png_do_gray_to_rgb(row_info, png_ptr->row_buf + 1);
#endif
#if defined(PNG_READ_BACKGROUND_SUPPORTED) || \
defined(PNG_READ_ALPHA_MODE_SUPPORTED)
if (png_ptr->transformations & PNG_COMPOSE)
png_do_compose(row_info, png_ptr->row_buf + 1, png_ptr);
#endif
#ifdef PNG_READ_GAMMA_SUPPORTED
if ((png_ptr->transformations & PNG_GAMMA) &&
#ifdef PNG_READ_RGB_TO_GRAY_SUPPORTED
/* Because RGB_TO_GRAY does the gamma transform. */
!(png_ptr->transformations & PNG_RGB_TO_GRAY) &&
#endif
#if defined(PNG_READ_BACKGROUND_SUPPORTED) ||\
defined(PNG_READ_ALPHA_MODE_SUPPORTED)
/* Because PNG_COMPOSE does the gamma transform if there is something to
* do (if there is an alpha channel or transparency.)
*/
!((png_ptr->transformations & PNG_COMPOSE) &&
((png_ptr->num_trans != 0) ||
(png_ptr->color_type & PNG_COLOR_MASK_ALPHA))) &&
#endif
/* Because png_init_read_transformations transforms the palette, unless
* RGB_TO_GRAY will do the transform.
*/
(png_ptr->color_type != PNG_COLOR_TYPE_PALETTE))
png_do_gamma(row_info, png_ptr->row_buf + 1, png_ptr);
#endif
#ifdef PNG_READ_STRIP_ALPHA_SUPPORTED
if ((png_ptr->transformations & PNG_STRIP_ALPHA) &&
(png_ptr->transformations & PNG_COMPOSE) &&
(row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA ||
row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA))
png_do_strip_channel(row_info, png_ptr->row_buf + 1,
0 /* at_start == false, because SWAP_ALPHA happens later */);
#endif
#ifdef PNG_READ_ALPHA_MODE_SUPPORTED
if ((png_ptr->transformations & PNG_ENCODE_ALPHA) &&
(row_info->color_type & PNG_COLOR_MASK_ALPHA))
png_do_encode_alpha(row_info, png_ptr->row_buf + 1, png_ptr);
#endif
#ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
if (png_ptr->transformations & PNG_SCALE_16_TO_8)
png_do_scale_16_to_8(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_STRIP_16_TO_8_SUPPORTED
/* There is no harm in doing both of these because only one has any effect,
* by putting the 'scale' option first if the app asks for scale (either by
* calling the API or in a TRANSFORM flag) this is what happens.
*/
if (png_ptr->transformations & PNG_16_TO_8)
png_do_chop(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_QUANTIZE_SUPPORTED
if (png_ptr->transformations & PNG_QUANTIZE)
{
png_do_quantize(row_info, png_ptr->row_buf + 1,
png_ptr->palette_lookup, png_ptr->quantize_index);
if (row_info->rowbytes == 0)
png_error(png_ptr, "png_do_quantize returned rowbytes=0");
}
#endif /* PNG_READ_QUANTIZE_SUPPORTED */
#ifdef PNG_READ_EXPAND_16_SUPPORTED
/* Do the expansion now, after all the arithmetic has been done. Notice
* that previous transformations can handle the PNG_EXPAND_16 flag if this
* is efficient (particularly true in the case of gamma correction, where
* better accuracy results faster!)
*/
if (png_ptr->transformations & PNG_EXPAND_16)
png_do_expand_16(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_GRAY_TO_RGB_SUPPORTED
/* NOTE: moved here in 1.5.4 (from much later in this list.) */
if ((png_ptr->transformations & PNG_GRAY_TO_RGB) &&
(png_ptr->flags & PNG_FLAG_BACKGROUND_IS_GRAY))
png_do_gray_to_rgb(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_INVERT_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_MONO)
png_do_invert(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_SHIFT_SUPPORTED
if (png_ptr->transformations & PNG_SHIFT)
png_do_unshift(row_info, png_ptr->row_buf + 1,
&(png_ptr->shift));
#endif
#ifdef PNG_READ_PACK_SUPPORTED
if (png_ptr->transformations & PNG_PACK)
png_do_unpack(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_CHECK_FOR_INVALID_INDEX_SUPPORTED
/* Added at libpng-1.5.10 */
if (row_info->color_type == PNG_COLOR_TYPE_PALETTE &&
png_ptr->num_palette_max >= 0)
png_do_check_palette_indexes(png_ptr, row_info);
#endif
#ifdef PNG_READ_BGR_SUPPORTED
if (png_ptr->transformations & PNG_BGR)
png_do_bgr(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_PACKSWAP_SUPPORTED
if (png_ptr->transformations & PNG_PACKSWAP)
png_do_packswap(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_FILLER_SUPPORTED
if (png_ptr->transformations & PNG_FILLER)
png_do_read_filler(row_info, png_ptr->row_buf + 1,
(png_uint_32)png_ptr->filler, png_ptr->flags);
#endif
#ifdef PNG_READ_INVERT_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_ALPHA)
png_do_read_invert_alpha(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_SWAP_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_ALPHA)
png_do_read_swap_alpha(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_READ_16BIT_SUPPORTED
#ifdef PNG_READ_SWAP_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_BYTES)
png_do_swap(row_info, png_ptr->row_buf + 1);
#endif
#endif
#ifdef PNG_READ_USER_TRANSFORM_SUPPORTED
if (png_ptr->transformations & PNG_USER_TRANSFORM)
{
if (png_ptr->read_user_transform_fn != NULL)
(*(png_ptr->read_user_transform_fn)) /* User read transform function */
(png_ptr, /* png_ptr */
row_info, /* row_info: */
/* png_uint_32 width; width of row */
/* png_size_t rowbytes; number of bytes in row */
/* png_byte color_type; color type of pixels */
/* png_byte bit_depth; bit depth of samples */
/* png_byte channels; number of channels (1-4) */
/* png_byte pixel_depth; bits per pixel (depth*channels) */
png_ptr->row_buf + 1); /* start of pixel data for row */
#ifdef PNG_USER_TRANSFORM_PTR_SUPPORTED
if (png_ptr->user_transform_depth)
row_info->bit_depth = png_ptr->user_transform_depth;
if (png_ptr->user_transform_channels)
row_info->channels = png_ptr->user_transform_channels;
#endif
row_info->pixel_depth = (png_byte)(row_info->bit_depth *
row_info->channels);
row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth, row_info->width);
}
#endif
}
#endif /* PNG_MNG_FEATURES_SUPPORTED */
#endif /* PNG_READ_TRANSFORMS_SUPPORTED */
#endif /* PNG_READ_SUPPORTED */

View File

@@ -257,6 +257,10 @@ png_crc_error(png_structrp png_ptr)
return (0);
}
#if defined(PNG_READ_iCCP_SUPPORTED) || defined(PNG_READ_iTXt_SUPPORTED) ||\
defined(PNG_READ_pCAL_SUPPORTED) || defined(PNG_READ_sCAL_SUPPORTED) ||\
defined(PNG_READ_sPLT_SUPPORTED) || defined(PNG_READ_tEXt_SUPPORTED) ||\
defined(PNG_READ_zTXt_SUPPORTED) || defined(PNG_SEQUENTIAL_READ_SUPPORTED)
/* Manage the read buffer; this simply reallocates the buffer if it is not small
* enough (or if it is not allocated). The routine returns a pointer to the
* buffer; if an error occurs and 'warn' is set the routine returns NULL, else
@@ -304,6 +308,7 @@ png_read_buffer(png_structrp png_ptr, png_alloc_size_t new_size, int warn)
return buffer;
}
#endif /* PNG_READ_iCCP|iTXt|pCAL|sCAL|sPLT|tEXt|zTXt|SEQUENTIAL_READ */
/* png_inflate_claim: claim the zstream for some nefarious purpose that involves
* decompression. Returns Z_OK on success, else a zlib error code. It checks
@@ -3789,7 +3794,6 @@ png_read_filter_row_paeth_1byte_pixel(png_row_infop row_info, png_bytep row,
/* Calculate the current pixel in a, and move the previous row pixel to c
* for the next time round the loop
*/
c = b;
a += *row;
*row++ = (png_byte)a;
}

View File

@@ -528,7 +528,7 @@ png_set_PLTE(png_structrp png_ptr, png_inforp info_ptr,
# endif
))
{
png_chunk_report(png_ptr, "Invalid palette", PNG_CHUNK_ERROR);
png_error(png_ptr, "Invalid palette");
return;
}

View File

@@ -569,7 +569,8 @@ png_debug_free(png_structp png_ptr, png_voidp ptr)
/* We must free the list element too, but first kill
the memory that is to be freed. */
memset(ptr, 0x55, pinfo->size);
png_free_default(png_ptr, pinfo);
if (pinfo)
free(pinfo);
pinfo = NULL;
break;
}
@@ -588,7 +589,8 @@ png_debug_free(png_structp png_ptr, png_voidp ptr)
if (verbose)
printf("Freeing %p\n", ptr);
png_free_default(png_ptr, ptr);
if (ptr)
free(ptr);
ptr = NULL;
}
#endif /* PNG_USER_MEM_SUPPORTED && PNG_DEBUG */
@@ -1574,14 +1576,12 @@ test_one_file(PNG_CONST char *inname, PNG_CONST char *outname)
#ifdef PNG_WRITE_SUPPORTED /* else nothing was written */
{
int wrote_question = 0;
for (;;)
{
static int wrote_question = 0;
png_size_t num_in, num_out;
char inbuf[256], outbuf[256];
num_in = fread(inbuf, 1, sizeof inbuf, fpin);
num_out = fread(outbuf, 1, sizeof outbuf, fpout);
@@ -1966,4 +1966,4 @@ main(void)
#endif
/* Generate a compiler error if there is an old png.h in the search path. */
typedef png_libpng_version_1_7_0beta19 Your_png_h_is_not_version_1_7_0beta19;
typedef png_libpng_version_1_7_0beta24 Your_png_h_is_not_version_1_7_0beta24;

View File

@@ -149,6 +149,8 @@ png_set_write_fn(png_structrp png_ptr, png_voidp io_ptr,
# else
png_ptr->output_flush_fn = output_flush_fn;
# endif
#else
PNG_UNUSED(output_flush_fn)
#endif /* PNG_WRITE_FLUSH_SUPPORTED */
#ifdef PNG_READ_SUPPORTED

View File

@@ -613,6 +613,71 @@ png_write_image(png_structrp png_ptr, png_bytepp image)
}
}
#ifdef PNG_MNG_FEATURES_SUPPORTED
/* Performs intrapixel differencing */
static void
png_do_write_intrapixel(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_write_intrapixel");
if ((row_info->color_type & PNG_COLOR_MASK_COLOR))
{
int bytes_per_pixel;
png_uint_32 row_width = row_info->width;
if (row_info->bit_depth == 8)
{
png_bytep rp;
png_uint_32 i;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 3;
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 4;
else
return;
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
*(rp) = (png_byte)((*rp - *(rp + 1)) & 0xff);
*(rp + 2) = (png_byte)((*(rp + 2) - *(rp + 1)) & 0xff);
}
}
#ifdef PNG_WRITE_16BIT_SUPPORTED
else if (row_info->bit_depth == 16)
{
png_bytep rp;
png_uint_32 i;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 6;
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 8;
else
return;
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
png_uint_32 s0 = (*(rp ) << 8) | *(rp + 1);
png_uint_32 s1 = (*(rp + 2) << 8) | *(rp + 3);
png_uint_32 s2 = (*(rp + 4) << 8) | *(rp + 5);
png_uint_32 red = (png_uint_32)((s0 - s1) & 0xffffL);
png_uint_32 blue = (png_uint_32)((s2 - s1) & 0xffffL);
*(rp ) = (png_byte)((red >> 8) & 0xff);
*(rp + 1) = (png_byte)(red & 0xff);
*(rp + 4) = (png_byte)((blue >> 8) & 0xff);
*(rp + 5) = (png_byte)(blue & 0xff);
}
}
#endif /* PNG_WRITE_16BIT_SUPPORTED */
}
}
#endif /* PNG_MNG_FEATURES_SUPPORTED */
/* Called by user to write a row of image data */
void PNGAPI
png_write_row(png_structrp png_ptr, png_const_bytep row)
@@ -1624,14 +1689,16 @@ png_write_image_16bit(png_voidp argument)
if (image->format & PNG_FORMAT_FLAG_ALPHA)
{
if (image->format & PNG_FORMAT_FLAG_AFIRST)
{
aindex = -1;
++input_row; /* To point to the first component */
++output_row;
}
# ifdef PNG_SIMPLIFIED_WRITE_AFIRST_SUPPORTED
if (image->format & PNG_FORMAT_FLAG_AFIRST)
{
aindex = -1;
++input_row; /* To point to the first component */
++output_row;
}
else
else
# endif
aindex = channels;
}
@@ -1780,14 +1847,16 @@ png_write_image_8bit(png_voidp argument)
png_bytep row_end;
int aindex;
if (image->format & PNG_FORMAT_FLAG_AFIRST)
{
aindex = -1;
++input_row; /* To point to the first component */
++output_row;
}
# ifdef PNG_SIMPLIFIED_WRITE_AFIRST_SUPPORTED
if (image->format & PNG_FORMAT_FLAG_AFIRST)
{
aindex = -1;
++input_row; /* To point to the first component */
++output_row;
}
else
else
# endif
aindex = channels;
/* Use row_end in place of a loop counter: */
@@ -1867,7 +1936,8 @@ png_image_set_PLTE(png_image_write_control *display)
const png_uint_32 format = image->format;
const int channels = PNG_IMAGE_SAMPLE_CHANNELS(format);
# ifdef PNG_FORMAT_BGR_SUPPORTED
# if defined(PNG_FORMAT_BGR_SUPPORTED) &&\
defined(PNG_SIMPLIFIED_WRITE_AFIRST_SUPPORTED)
const int afirst = (format & PNG_FORMAT_FLAG_AFIRST) != 0 &&
(format & PNG_FORMAT_FLAG_ALPHA) != 0;
# else

View File

@@ -14,90 +14,14 @@
#include "pngpriv.h"
#ifdef PNG_WRITE_SUPPORTED
#ifdef PNG_WRITE_TRANSFORMS_SUPPORTED
/* Transform the data according to the user's wishes. The order of
* transformations is significant.
*/
void /* PRIVATE */
png_do_write_transformations(png_structrp png_ptr, png_row_infop row_info)
{
png_debug(1, "in png_do_write_transformations");
if (png_ptr == NULL)
return;
#ifdef PNG_WRITE_USER_TRANSFORM_SUPPORTED
if (png_ptr->transformations & PNG_USER_TRANSFORM)
if (png_ptr->write_user_transform_fn != NULL)
(*(png_ptr->write_user_transform_fn)) /* User write transform
function */
(png_ptr, /* png_ptr */
row_info, /* row_info: */
/* png_uint_32 width; width of row */
/* png_size_t rowbytes; number of bytes in row */
/* png_byte color_type; color type of pixels */
/* png_byte bit_depth; bit depth of samples */
/* png_byte channels; number of channels (1-4) */
/* png_byte pixel_depth; bits per pixel (depth*channels) */
png_ptr->row_buf + 1); /* start of pixel data for row */
#endif
#ifdef PNG_WRITE_FILLER_SUPPORTED
if (png_ptr->transformations & PNG_FILLER)
png_do_strip_channel(row_info, png_ptr->row_buf + 1,
!(png_ptr->flags & PNG_FLAG_FILLER_AFTER));
#endif
#ifdef PNG_WRITE_PACKSWAP_SUPPORTED
if (png_ptr->transformations & PNG_PACKSWAP)
png_do_packswap(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_WRITE_PACK_SUPPORTED
if (png_ptr->transformations & PNG_PACK)
png_do_pack(row_info, png_ptr->row_buf + 1,
(png_uint_32)png_ptr->bit_depth);
#endif
#ifdef PNG_WRITE_SWAP_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_BYTES)
png_do_swap(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_WRITE_SHIFT_SUPPORTED
if (png_ptr->transformations & PNG_SHIFT)
png_do_shift(row_info, png_ptr->row_buf + 1,
&(png_ptr->shift));
#endif
#ifdef PNG_WRITE_SWAP_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_ALPHA)
png_do_write_swap_alpha(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_WRITE_INVERT_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_ALPHA)
png_do_write_invert_alpha(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_WRITE_BGR_SUPPORTED
if (png_ptr->transformations & PNG_BGR)
png_do_bgr(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_WRITE_INVERT_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_MONO)
png_do_invert(row_info, png_ptr->row_buf + 1);
#endif
}
#ifdef PNG_WRITE_PACK_SUPPORTED
/* Pack pixels into bytes. Pass the true bit depth in bit_depth. The
* row_info bit depth should be 8 (one pixel per byte). The channels
* should be 1 (this only happens on grayscale and paletted images).
*/
void /* PRIVATE */
static void
png_do_pack(png_row_infop row_info, png_bytep row, png_uint_32 bit_depth)
{
png_debug(1, "in png_do_pack");
@@ -242,7 +166,7 @@ png_do_pack(png_row_infop row_info, png_bytep row, png_uint_32 bit_depth)
* would pass 3 as bit_depth, and this routine would translate the
* data to 0 to 15.
*/
void /* PRIVATE */
static void
png_do_shift(png_row_infop row_info, png_bytep row,
png_const_color_8p bit_depth)
{
@@ -381,7 +305,7 @@ png_do_shift(png_row_infop row_info, png_bytep row,
#endif
#ifdef PNG_WRITE_SWAP_ALPHA_SUPPORTED
void /* PRIVATE */
static void
png_do_write_swap_alpha(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_write_swap_alpha");
@@ -475,7 +399,7 @@ png_do_write_swap_alpha(png_row_infop row_info, png_bytep row)
#endif
#ifdef PNG_WRITE_INVERT_ALPHA_SUPPORTED
void /* PRIVATE */
static void
png_do_write_invert_alpha(png_row_infop row_info, png_bytep row)
{
png_debug(1, "in png_do_write_invert_alpha");
@@ -568,70 +492,81 @@ png_do_write_invert_alpha(png_row_infop row_info, png_bytep row)
}
}
#endif
#endif /* PNG_WRITE_TRANSFORMS_SUPPORTED */
#ifdef PNG_MNG_FEATURES_SUPPORTED
/* Undoes intrapixel differencing */
/* Transform the data according to the user's wishes. The order of
* transformations is significant.
*/
void /* PRIVATE */
png_do_write_intrapixel(png_row_infop row_info, png_bytep row)
png_do_write_transformations(png_structrp png_ptr, png_row_infop row_info)
{
png_debug(1, "in png_do_write_intrapixel");
png_debug(1, "in png_do_write_transformations");
if ((row_info->color_type & PNG_COLOR_MASK_COLOR))
{
int bytes_per_pixel;
png_uint_32 row_width = row_info->width;
if (row_info->bit_depth == 8)
{
png_bytep rp;
png_uint_32 i;
if (png_ptr == NULL)
return;
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 3;
#ifdef PNG_WRITE_USER_TRANSFORM_SUPPORTED
if (png_ptr->transformations & PNG_USER_TRANSFORM)
if (png_ptr->write_user_transform_fn != NULL)
(*(png_ptr->write_user_transform_fn)) /* User write transform
function */
(png_ptr, /* png_ptr */
row_info, /* row_info: */
/* png_uint_32 width; width of row */
/* png_size_t rowbytes; number of bytes in row */
/* png_byte color_type; color type of pixels */
/* png_byte bit_depth; bit depth of samples */
/* png_byte channels; number of channels (1-4) */
/* png_byte pixel_depth; bits per pixel (depth*channels) */
png_ptr->row_buf + 1); /* start of pixel data for row */
#endif
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 4;
#ifdef PNG_WRITE_FILLER_SUPPORTED
if (png_ptr->transformations & PNG_FILLER)
png_do_strip_channel(row_info, png_ptr->row_buf + 1,
!(png_ptr->flags & PNG_FLAG_FILLER_AFTER));
#endif
else
return;
#ifdef PNG_WRITE_PACKSWAP_SUPPORTED
if (png_ptr->transformations & PNG_PACKSWAP)
png_do_packswap(row_info, png_ptr->row_buf + 1);
#endif
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
*(rp) = (png_byte)((*rp - *(rp + 1)) & 0xff);
*(rp + 2) = (png_byte)((*(rp + 2) - *(rp + 1)) & 0xff);
}
}
#ifdef PNG_WRITE_PACK_SUPPORTED
if (png_ptr->transformations & PNG_PACK)
png_do_pack(row_info, png_ptr->row_buf + 1,
(png_uint_32)png_ptr->bit_depth);
#endif
#ifdef PNG_WRITE_16BIT_SUPPORTED
else if (row_info->bit_depth == 16)
{
png_bytep rp;
png_uint_32 i;
#ifdef PNG_WRITE_SWAP_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_BYTES)
png_do_swap(row_info, png_ptr->row_buf + 1);
#endif
if (row_info->color_type == PNG_COLOR_TYPE_RGB)
bytes_per_pixel = 6;
#ifdef PNG_WRITE_SHIFT_SUPPORTED
if (png_ptr->transformations & PNG_SHIFT)
png_do_shift(row_info, png_ptr->row_buf + 1,
&(png_ptr->shift));
#endif
else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
bytes_per_pixel = 8;
#ifdef PNG_WRITE_SWAP_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_SWAP_ALPHA)
png_do_write_swap_alpha(row_info, png_ptr->row_buf + 1);
#endif
else
return;
#ifdef PNG_WRITE_INVERT_ALPHA_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_ALPHA)
png_do_write_invert_alpha(row_info, png_ptr->row_buf + 1);
#endif
for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
{
png_uint_32 s0 = (*(rp ) << 8) | *(rp + 1);
png_uint_32 s1 = (*(rp + 2) << 8) | *(rp + 3);
png_uint_32 s2 = (*(rp + 4) << 8) | *(rp + 5);
png_uint_32 red = (png_uint_32)((s0 - s1) & 0xffffL);
png_uint_32 blue = (png_uint_32)((s2 - s1) & 0xffffL);
*(rp ) = (png_byte)((red >> 8) & 0xff);
*(rp + 1) = (png_byte)(red & 0xff);
*(rp + 4) = (png_byte)((blue >> 8) & 0xff);
*(rp + 5) = (png_byte)(blue & 0xff);
}
}
#endif /* PNG_WRITE_16BIT_SUPPORTED */
}
#ifdef PNG_WRITE_BGR_SUPPORTED
if (png_ptr->transformations & PNG_BGR)
png_do_bgr(row_info, png_ptr->row_buf + 1);
#endif
#ifdef PNG_WRITE_INVERT_SUPPORTED
if (png_ptr->transformations & PNG_INVERT_MONO)
png_do_invert(row_info, png_ptr->row_buf + 1);
#endif
}
#endif /* PNG_MNG_FEATURES_SUPPORTED */
#endif /* PNG_WRITE_TRANSFORMS_SUPPORTED */
#endif /* PNG_WRITE_SUPPORTED */

View File

@@ -1,7 +1,7 @@
VisualStudio instructions
libpng version 1.7.0beta19 - September 30, 2013
libpng version 1.7.0beta24 - December 15, 2013
Copyright (c) 1998-2010 Glenn Randers-Pehrson

View File

@@ -2,7 +2,7 @@
<!--
* zlib.props - location of zlib source
*
* libpng version 1.7.0beta19 - September 30, 2013
* libpng version 1.7.0beta24 - December 15, 2013
*
* Copyright (c) 1998-2011 Glenn Randers-Pehrson
*

View File

@@ -1,9 +1,9 @@
Makefiles for libpng version 1.7.0beta19 - September 30, 2013
Makefiles for libpng version 1.7.0beta24 - December 15, 2013
pnglibconf.h.prebuilt => Stores configuration settings
makefile.linux => Linux/ELF makefile
(gcc, creates libpng17.so.17.1.7.0beta19)
(gcc, creates libpng17.so.17.1.7.0beta24)
makefile.gcc => Generic makefile (gcc, creates static libpng.a)
makefile.knr => Archaic UNIX Makefile that converts files with
ansi2knr (Requires ansi2knr.c from
@@ -20,7 +20,7 @@ pnglibconf.h.prebuilt => Stores configuration settings
makefile.dec => DEC Alpha UNIX makefile
makefile.dj2 => DJGPP 2 makefile
makefile.elf => Linux/ELF makefile symbol versioning,
(gcc, creates libpng17.so.17.1.7.0beta19)
(gcc, creates libpng17.so.17.1.7.0beta24)
makefile.freebsd => FreeBSD makefile
makefile.gcc => Generic gcc makefile
makefile.hpgcc => HPUX makefile using gcc
@@ -35,12 +35,12 @@ pnglibconf.h.prebuilt => Stores configuration settings
makefile.os2 => OS/2 Makefile (gcc and emx, requires libpng.def)
makefile.sco => For SCO OSr5 ELF and Unixware 7 with Native cc
makefile.sggcc => Silicon Graphics (gcc,
creates libpng17.so.17.1.7.0beta19)
creates libpng17.so.17.1.7.0beta24)
makefile.sgi => Silicon Graphics IRIX makefile (cc, creates static lib)
makefile.solaris => Solaris 2.X makefile (gcc,
creates libpng17.so.17.1.7.0beta19)
creates libpng17.so.17.1.7.0beta24)
makefile.so9 => Solaris 9 makefile (gcc,
creates libpng17.so.17.1.7.0beta19)
creates libpng17.so.17.1.7.0beta24)
makefile.std => Generic UNIX makefile (cc, creates static libpng.a)
makefile.sunos => Sun makefile
makefile.32sunu => Sun Ultra 32-bit makefile

View File

@@ -21,7 +21,7 @@ PNG_DFN "OS2 DESCRIPTION "PNG image compression library""
PNG_DFN "OS2 CODE PRELOAD MOVEABLE DISCARDABLE"
PNG_DFN ""
PNG_DFN "EXPORTS"
PNG_DFN ";Version 1.7.0beta19"
PNG_DFN ";Version 1.7.0beta24"
#define PNG_EXPORTA(ordinal, type, name, args, attributes)\
PNG_DFN "@" SYMBOL_PREFIX "@@" name "@"

View File

@@ -11,7 +11,7 @@
# Modeled after libxml-config.
version=1.7.0beta19
version=1.7.0beta24
prefix=""
libdir=""
libs=""

View File

@@ -5,6 +5,6 @@ includedir=@includedir@/libpng17
Name: libpng
Description: Loads and saves PNG files
Version: 1.7.0beta19
Version: 1.7.0beta24
Libs: -L${libdir} -lpng17
Cflags: -I${includedir}

View File

@@ -17,7 +17,7 @@ INCSDIR=${LOCALBASE}/include/libpng17
LIB= png17
SHLIB_MAJOR= 0
SHLIB_MINOR= 1.7.0beta19
SHLIB_MINOR= 1.7.0beta24
SRCS= png.c pngset.c pngget.c pngrutil.c pngtrans.c pngwutil.c \
pngread.c pngrio.c pngwio.c pngwrite.c pngrtran.c \
pngwtran.c pngmem.c pngerror.c pngpread.c

View File

@@ -17,7 +17,7 @@ INCSDIR=${LOCALBASE}/include
LIB= png
SHLIB_MAJOR= 17
SHLIB_MINOR= 1.7.0beta19
SHLIB_MINOR= 1.7.0beta24
SRCS= png.c pngset.c pngget.c pngrutil.c pngtrans.c pngwutil.c \
pngread.c pngrio.c pngwio.c pngwrite.c pngrtran.c \
pngwtran.c pngmem.c pngerror.c pngpread.c

View File

@@ -11,7 +11,7 @@ LIBDIR= ${PREFIX}/lib
MANDIR= ${PREFIX}/man/cat
SHLIB_MAJOR= 17
SHLIB_MINOR= 1.7.0beta19
SHLIB_MINOR= 1.7.0beta24
LIB= png
SRCS= png.c pngerror.c pngget.c pngmem.c pngpread.c \

View File

@@ -282,7 +282,7 @@ $1 == "option" && NF >= 2{
for (i=istart; i<=NF; ++i) {
val=$(i)
sub(/,$/,"",val)
if (val == "on" || val == "off" || val == "disabled") {
if (val == "on" || val == "off" || val == "disabled" || val =="enabled") {
key = ""
if (onoff != val) {
# on or off can zap disabled or enabled:
@@ -687,9 +687,11 @@ END{
}
# if
have_ifs = 0
nreqs = split(iffs[i], r)
print "#undef PNG_no_if" >out
if (nreqs > 0) {
have_ifs = 1
print "/* if" iffs[i], "*/" >out
print "#define PNG_no_if 1" >out
for (j=1; j<=nreqs; ++j) {
@@ -727,7 +729,10 @@ END{
print "# ifndef PNG_" i "_SUPPORTED /*!command line*/" >out
print "# ifdef PNG_not_enabled /*!enabled*/" >out
if (option[i] == "off" || option[i] == "disabled" && everything != "on" || option[i] == "enabled" && everything == "off") {
# 'have_ifs' here means that everything = "off" still allows an 'if' on
# an otherwise enabled option to turn it on; otherwise the 'if'
# handling is effectively disabled by 'everything = off'
if (option[i] == "off" || option[i] == "disabled" && everything != "on" || option[i] == "enabled" && everything == "off" && !have_ifs) {
print "# undef PNG_on /*default off*/" >out
} else {
print "# ifdef PNG_NO_" i >out

View File

@@ -328,12 +328,16 @@ option BENIGN_READ_ERRORS requires BENIGN_ERRORS
option MNG_FEATURES
# Arithmetic options, the first is the big switch that chooses between internal
# floating and fixed point arithmetic implementations - it does not affect any
# floating and fixed point arithmetic implementations - it does not affect an
# APIs. The second two (the _POINT settings) switch off individual APIs.
#
# Prior to libpng 1.6.8 one of the API (_POINT) variants had to be selected. At
# 1.6.8 this restriction has been removed; the simplified API can be used
# without enabling any of the low level fixed/floating APIs.
option FLOATING_ARITHMETIC
option FLOATING_POINT enables ok_math
option FIXED_POINT enables ok_math
option FLOATING_POINT
option FIXED_POINT
# The following is always on (defined empty)
@@ -461,7 +465,7 @@ option READ_TRANSFORMS requires READ
# If you handle gamma issues outside libpng then you do not need the libpng
# gamma processing; and it is an enormous waste of space. You just need to
# remove the use of libpng APIs that depend on it.
option READ_GAMMA requires READ_TRANSFORMS, READ_gAMA
option READ_GAMMA requires READ_TRANSFORMS, READ_gAMA, READ_sRGB
option READ_ALPHA_MODE requires READ_TRANSFORMS, READ_GAMMA
option READ_BACKGROUND requires READ_TRANSFORMS, READ_STRIP_ALPHA, READ_GAMMA
@@ -474,7 +478,7 @@ option READ_INVERT_ALPHA requires READ_TRANSFORMS
option READ_INVERT requires READ_TRANSFORMS
option READ_PACK requires READ_TRANSFORMS
option READ_PACKSWAP requires READ_TRANSFORMS
option READ_RGB_TO_GRAY requires READ_TRANSFORMS, READ_GAMMA
option READ_RGB_TO_GRAY requires READ_TRANSFORMS, READ_GAMMA enables COLORSPACE
option READ_SCALE_16_TO_8 requires READ_TRANSFORMS
option READ_SHIFT requires READ_TRANSFORMS
option READ_STRIP_16_TO_8 requires READ_TRANSFORMS
@@ -502,6 +506,7 @@ option INCH_CONVERSIONS
= INCH_CONVERSIONS INCH_CONVERSIONS
# API to build a grayscale palette
# NOTE: this is not used internally by libpng at present.
option BUILD_GRAYSCALE_PALETTE
@@ -831,13 +836,13 @@ option SAVE_INT_32 disabled
option WRITE_OPTIMIZE_CMF requires WRITE
option READ_COMPRESSED_TEXT disabled
option READ_iCCP enables READ_COMPRESSED_TEXT
option READ_iTXt enables READ_COMPRESSED_TEXT
option READ_zTXt enables READ_COMPRESSED_TEXT
option READ_COMPRESSED_TEXT enables READ_TEXT
option WRITE_oFFs enables SAVE_INT_32
option WRITE_pCAL enables SAVE_INT_32
option WRITE_cHRM enables SAVE_INT_32
option WRITE_COMPRESSED_TEXT disabled
option WRITE_iCCP enables WRITE_COMPRESSED_TEXT
@@ -860,35 +865,49 @@ option READ_CHECK_FOR_INVALID_INDEX requires READ disabled
option WRITE_CHECK_FOR_INVALID_INDEX requires WRITE disabled
# added at libpng-1.5.15
option GET_PALETTE_MAX enables READ_GET_PALETTE_MAX, WRITE_GET_PALETTE_MAX
option GET_PALETTE_MAX enables READ_GET_PALETTE_MAX WRITE_GET_PALETTE_MAX
option READ_GET_PALETTE_MAX requires READ_CHECK_FOR_INVALID_INDEX disabled
option WRITE_GET_PALETTE_MAX requires WRITE_CHECK_FOR_INVALID_INDEX disabled
# Simplified API options (added at libpng-1.6.0)
# In libpng 1.6.8 the handling of these options was changed to used 'requires'
# throughout, so that disabling some of the low level support always disables
# the base simplified read/write API. This much simplifies the handling and
# makes 'everything = off' work in a more intuitive way. It eliminates a
# previously reported feature that APIs previously enabled by the simplified
# API couldn't be turned off without explicitly turning off the simplified
# APIs.
#
# Read:
option SIMPLIFIED_READ,
requires SEQUENTIAL_READ READ_TRANSFORMS, SETJMP, BENIGN_ERRORS READ_GAMMA,
enables READ_EXPAND, READ_16BIT READ_EXPAND_16, READ_SCALE_16_TO_8,
READ_RGB_TO_GRAY, READ_ALPHA_MODE READ_BACKGROUND READ_STRIP_ALPHA,
READ_FILLER, READ_SWAP
requires SEQUENTIAL_READ, READ_TRANSFORMS, SETJMP, BENIGN_ERRORS,
READ_EXPAND, READ_16BIT, READ_EXPAND_16, READ_SCALE_16_TO_8,
READ_RGB_TO_GRAY, READ_ALPHA_MODE, READ_BACKGROUND, READ_STRIP_ALPHA,
READ_FILLER, READ_SWAP, READ_PACK, READ_GRAY_TO_RGB, READ_GAMMA,
READ_tRNS, READ_bKGD, READ_gAMA, READ_cHRM, READ_sRGB, READ_sBIT
option SIMPLIFIED_READ_AFIRST requires SIMPLIFIED_READ disabled
option READ_SWAP_ALPHA enables SIMPLIFIED_READ_AFIRST
# AFIRST and BGR read options:
# Prior to libpng 1.6.8 these were disabled but switched on if the low level
# libpng routines that do the swaps were enabled. This worked but was
# confusing. In libpng 1.6.8 the options were changed to simple 'requires'
# and are enabled by default. This should work the same way in practice.
option SIMPLIFIED_READ_AFIRST enables FORMAT_AFIRST,
requires SIMPLIFIED_READ READ_SWAP_ALPHA
option SIMPLIFIED_READ_BGR requires SIMPLIFIED_READ disabled
option READ_BGR enables SIMPLIFIED_READ_BGR
option SIMPLIFIED_READ_BGR enables FORMAT_BGR,
requires SIMPLIFIED_READ READ_BGR
# Write:
option SIMPLIFIED_WRITE,
requires WRITE STDIO, SETJMP,
enables WRITE_SWAP WRITE_gAMA, WRITE_sRGB WRITE_cHRM
requires WRITE STDIO, SETJMP, WRITE_SWAP, WRITE_PACK,
WRITE_tRNS, WRITE_gAMA, WRITE_sRGB, WRITE_cHRM
option SIMPLIFIED_WRITE_AFIRST requires SIMPLIFIED_WRITE disabled
option WRITE_SWAP_ALPHA enables SIMPLIFIED_WRITE_AFIRST
option SIMPLIFIED_WRITE_AFIRST enables FORMAT_AFIRST,
requires SIMPLIFIED_WRITE WRITE_SWAP_ALPHA
option SIMPLIFIED_WRITE_BGR requires SIMPLIFIED_WRITE disabled
option WRITE_BGR enables SIMPLIFIED_WRITE_BGR
option SIMPLIFIED_WRITE_BGR enables FORMAT_BGR,
requires SIMPLIFIED_WRITE WRITE_BGR
# Formats:
option FORMAT_AFIRST if SIMPLIFIED_READ_AFIRST, SIMPLIFIED_WRITE_AFIRST
option FORMAT_BGR if SIMPLIFIED_READ_BGR, SIMPLIFIED_WRITE_BGR
option FORMAT_AFIRST disabled
option FORMAT_BGR disabled

View File

@@ -1,8 +1,8 @@
/* libpng 1.7.0beta19 STANDARD API DEFINITION */
/* libpng 1.7.0beta24 STANDARD API DEFINITION */
/* pnglibconf.h - library build configuration */
/* Libpng version 1.7.0beta19 - September 30, 2013 */
/* Libpng version 1.7.0beta24 - December 15, 2013 */
/* Copyright (c) 1998-2013 Glenn Randers-Pehrson */

View File

@@ -1,4 +1,4 @@
;Version 1.7.0beta19
;Version 1.7.0beta24
;--------------------------------------------------------------
; LIBPNG symbol list as a Win32 DEF file
; Contains all the symbols that can be exported from libpng