Make Visual C++ happy

git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@465 632fc199-4ca6-4c93-a231-07263d6284db
Fix compiler warnings
2011-02-27 00:07:51 +00:00 · 2011-02-26 21:08:39 +00:00 · 2011-02-26 21:07:50 +00:00 · 2011-02-26 19:46:27 +00:00 · 2011-02-25 06:21:15 +00:00 · 2011-02-25 02:13:07 +00:00
31 changed files with 675 additions and 503 deletions
--- a/BUILDING.txt
+++ b/BUILDING.txt
@@ -91,17 +91,17 @@ This will generate the following files under .libs/
      Development stub for TurboJPEG/OSS shared library

 {version} is 62.0.0, 7.0.0, or 8.0.2, depending on whether libjpeg v6b
-(default), v7, or v8b emulation is enabled.  If using Cygwin, {version} is
+(default), v7, or v8 emulation is enabled.  If using Cygwin, {version} is
 62, 7, or 8.


-libjpeg v7 or v8b Emulation
---------------------------
+libjpeg v7 or v8 Emulation
+--------------------------

 Add --with-jpeg7 to the configure command line to build a version of
 libjpeg-turbo that is compatible with libjpeg v7.  Add --with-jpeg8 to the
 configure command to build a version of libjpeg-turbo that is compatible with
-libjpeg v8b.  See README-turbo.txt for more information on libjpeg v7 and v8b
+libjpeg v8.  See README-turbo.txt for more information on libjpeg v7 and v8
 emulation.


@@ -110,7 +110,7 @@ Arithmetic Coding Support

 Since the patent on arithmetic coding has expired, this functionality has been
 included in this release of libjpeg-turbo.  libjpeg-turbo's implementation is
-based on the implementation in libjpeg v8b, but it works when emulating libjpeg
+based on the implementation in libjpeg v8, but it works when emulating libjpeg
 v7 or v6b as well.  The default is to enable both arithmetic encoding and
 decoding, but those who have philosophical objections to arithmetic coding can
 add --without-arith-enc or --without-arith-dec to the configure command line to
@@ -154,7 +154,7 @@ Build Recipes

 Add

-  --host i686-pc-linux-gnu CFLAGS='-O3 -m32' CXXFLAGS='-O3 -m32' LDFLAGS=-m32
+  --host i686-pc-linux-gnu CFLAGS='-O3 -m32' LDFLAGS=-m32

 to the configure command line.

@@ -175,7 +175,7 @@ installed.

 Add

-  CFLAGS='-O3 -m32' CXXFLAGS='-O3 -m32' LDFLAGS=-m32
+  CFLAGS='-O3 -m32' LDFLAGS=-m32

 to the configure command line.

@@ -187,8 +187,6 @@ Add

  --host x86_64-apple-darwin NASM=/opt/local/bin/nasm \
  CFLAGS='-isysroot /Developer/SDKs/MacOSX10.4u.sdk \
-    -mmacosx-version-min=10.4 -O3' \
-    CXXFLAGS='-isysroot /Developer/SDKs/MacOSX10.4u.sdk \
    -mmacosx-version-min=10.4 -O3' \
    LDFLAGS='-isysroot /Developer/SDKs/MacOSX10.4u.sdk \
    -mmacosx-version-min=10.4'
@@ -203,8 +201,6 @@ MacPorts, must be installed.
 Add

  CFLAGS='-isysroot /Developer/SDKs/MacOSX10.4u.sdk \
-    -mmacosx-version-min=10.4 -O3 -m32' \
-    CXXFLAGS='-isysroot /Developer/SDKs/MacOSX10.4u.sdk \
    -mmacosx-version-min=10.4 -O3 -m32' \
    LDFLAGS='-isysroot /Developer/SDKs/MacOSX10.4u.sdk \
    -mmacosx-version-min=10.4 -m32'
@@ -217,7 +213,7 @@ to the configure command line.  The OS X 10.4 SDK must be installed.

 Add

-  --host x86_64-pc-solaris CFLAGS='-O3 -m64' CXXFLAGS='-O3 -m64' LDFLAGS=-m64
+  --host x86_64-pc-solaris CFLAGS='-O3 -m64' LDFLAGS=-m64

 to the configure command line.

@@ -227,8 +223,8 @@ to the configure command line.

 Add

-  --host i386-unknown-freebsd CC='gcc -B /usr/lib32' CXX='g++ -B/usr/lib32' \
-    CFLAGS='-O3 -m32' CXXFLAGS='-O3 -m32' LDFLAGS='-B/usr/lib32'
+  --host i386-unknown-freebsd CC='gcc -B /usr/lib32' CFLAGS='-O3 -m32' \
+    LDFLAGS='-B/usr/lib32'

 to the configure command line.  NASM 2.07 or later from FreeBSD ports must be
 installed.
@@ -239,15 +235,14 @@ Sun Studio

 Add

-  CC=cc CXX=CC
+  CC=cc

 to the configure command line.  libjpeg-turbo will automatically be built with
-the maximum optimization level (-xO5) unless you override CFLAGS and CXXFLAGS.
+the maximum optimization level (-xO5) unless you override CFLAGS.

 To build a 64-bit version of libjpeg-turbo using Sun Studio, add

-  --host x86_64-pc-solaris CC=cc CXX=CC CFLAGS='-xO5 -m64' \
-    CXXFLAGS='-xO5 -m64' LDFLAGS=-m64
+  --host x86_64-pc-solaris CC=cc CFLAGS='-xO5 -m64' LDFLAGS=-m64

 to the configure command line.

@@ -346,7 +341,7 @@ The following files will be generated under {build_directory}:
      Import library for TurboJPEG/OSS DLL

 {version} is 62, 7, or 8, depending on whether libjpeg v6b (default), v7, or
-v8b emulation is enabled.
+v8 emulation is enabled.


 Visual C++ (IDE)
@@ -380,7 +375,7 @@ This will generate the following files under {build_directory}:

 {configuration} is Debug, Release, RelWithDebInfo, or MinSizeRel, depending on
 the configuration you built in the IDE, and {version} is 62, 7, or 8,
-depending on whether libjpeg v6b (default), v7, or v8b emulation is enabled.
+depending on whether libjpeg v6b (default), v7, or v8 emulation is enabled.


 MinGW
@@ -406,7 +401,7 @@ This will generate the following files under {build_directory}
      Import library for TurboJPEG/OSS DLL

 {version} is 62, 7, or 8, depending on whether libjpeg v6b (default), v7, or
-v8b emulation is enabled.
+v8 emulation is enabled.


 Debug Build
@@ -417,13 +412,13 @@ NMake, remove "-DCMAKE_BUILD_TYPE=Release" (Debug builds are the default with
 NMake.)


-libjpeg v7 or v8b Emulation
---------------------------
+libjpeg v7 or v8 Emulation
+--------------------------

 Add "-DWITH_JPEG7=1" to the cmake command line to build a version of
 libjpeg-turbo that is compatible with libjpeg v7.  Add "-DWITH_JPEG8=1" to the
 cmake command to build a version of libjpeg-turbo that is compatible with
-libjpeg v8b.  See README-turbo.txt for more information on libjpeg v7 and v8b
+libjpeg v8.  See README-turbo.txt for more information on libjpeg v7 and v8
 emulation.


@@ -432,7 +427,7 @@ Arithmetic Coding Support

 Since the patent on arithmetic coding has expired, this functionality has been
 included in this release of libjpeg-turbo.  libjpeg-turbo's implementation is
-based on the implementation in libjpeg v8b, but it works when emulating libjpeg
+based on the implementation in libjpeg v8, but it works when emulating libjpeg
 v7 or v6b as well.  The default is to enable both arithmetic encoding and
 decoding, but those who have philosophical objections to arithmetic coding can
 add "-DWITH_ARITH_ENC=0" or "-DWITH_ARITH_DEC=0" to the cmake command line to
@@ -470,7 +465,7 @@ Build Recipes
 ----------------------------

  cd {build_directory}
-  CC=/usr/bin/x86_64-w64-mingw32-gcc CXX=/usr/bin/x86_64-w64-mingw32-g++ \
+  CC=/usr/bin/x86_64-w64-mingw32-gcc \
    cmake -G "Unix Makefiles" -DCMAKE_SYSTEM_NAME=Windows \
    -DCMAKE_AR=/usr/bin/x86_64-w64-mingw32-ar \
    -DCMAKE_RANLIB=/usr/bin/x86_64-w64-mingw32-ranlib {source_directory}
@@ -485,7 +480,7 @@ mingw64-x86_64-gcc-g++ packages (and their dependencies) must be installed.
 ----------------------------

  cd {build_directory}
-  CC=/usr/bin/i686-w64-mingw32-gcc CXX=/usr/bin/i686-w64-mingw32-g++ \
+  CC=/usr/bin/i686-w64-mingw32-gcc \
    cmake -G "Unix Makefiles" -DCMAKE_SYSTEM_NAME=Windows \
    -DDCMAKE_AR=/usr/bin/i686-w64-mingw32-ar \
    -DCMAKE_RANLIB=/usr/bin/i686-w64-mingw32-ranlib {source_directory}
@@ -504,7 +499,6 @@ toolchain (which is faster than the Cygwin version):

  cd {build_directory}
  CC={mingw-w64_binary_path}/x86_64-w64-mingw32-gcc \
-    CXX={mingw-w64_binary_path}/x86_64-w64-mingw32-g++ \
    cmake -G "MSYS Makefiles" \
    -DCMAKE_AR={mingw-w64_binary_path}/x86_64-w64-mingw32-ar \
    -DCMAKE_RANLIB={mingw-w64_binary_path}/x86_64-w64-mingw32-ranlib \
@@ -517,7 +511,6 @@ MinGW Build on Linux

  cd {build_directory}
  CC={mingw_binary_path}/i386-mingw32-gcc \
-    CXX={mingw_binary_path}/i386-mingw32-g++ \
    cmake -G "Unix Makefiles" -DCMAKE_SYSTEM_NAME=Windows \
    -DCMAKE_AR={mingw_binary_path}/i386-mingw32-ar \
    -DCMAKE_RANLIB={mingw_binary_path}/i386-mingw32-ranlib \
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -4,8 +4,8 @@

 cmake_minimum_required(VERSION 2.6)

-project(libjpeg-turbo)
-set(VERSION 1.0.90)
+project(libjpeg-turbo C)
+set(VERSION 1.1.0)

 if(MINGW OR CYGWIN)
  execute_process(COMMAND "date" "+%Y%m%d" OUTPUT_VARIABLE BUILD)
@@ -65,7 +65,7 @@ if(WITH_JPEG8)
  set(JPEG_LIB_VERSION 80)
  set(DLL_VERSION 8)
  set(FULLVERSION ${DLL_VERSION}.0.2)
-  message(STATUS "Emulating libjpeg v8b API/ABI")
+  message(STATUS "Emulating libjpeg v8 API/ABI")
 elseif(WITH_JPEG7)
  set(JPEG_LIB_VERSION 70)
  set(DLL_VERSION 7)
@@ -76,9 +76,7 @@ endif(WITH_JPEG8)
 if(MSVC)
  # Use the static C library for all build types
  foreach(var CMAKE_C_FLAGS CMAKE_C_FLAGS_DEBUG CMAKE_C_FLAGS_RELEASE
-    CMAKE_C_FLAGS_MINSIZEREL CMAKE_C_FLAGS_RELWITHDEBINFO
-    CMAKE_CXX_FLAGS CMAKE_CXX_FLAGS_DEBUG CMAKE_CXX_FLAGS_RELEASE
-    CMAKE_CXX_FLAGS_MINSIZEREL CMAKE_CXX_FLAGS_RELWITHDEBINFO)
+    CMAKE_C_FLAGS_MINSIZEREL CMAKE_C_FLAGS_RELWITHDEBINFO)
    if(${var} MATCHES "/MD")
      string(REGEX REPLACE "/MD" "/MT" ${var} "${${var}}")
    endif()
@@ -177,10 +175,10 @@ target_link_libraries(jpegut turbojpeg)
 add_executable(jpegut-static jpegut.c)
 target_link_libraries(jpegut-static turbojpeg-static)

-add_executable(jpgtest jpgtest.cxx bmp.c)
+add_executable(jpgtest jpgtest.c bmp.c)
 target_link_libraries(jpgtest turbojpeg)

-add_executable(jpgtest-static jpgtest.cxx bmp.c)
+add_executable(jpgtest-static jpgtest.c bmp.c)
 target_link_libraries(jpgtest-static turbojpeg-static)

 add_executable(cjpeg-static cjpeg.c cdjpeg.c rdbmp.c rdgif.c rdppm.c rdswitch.c
@@ -214,6 +212,8 @@ add_test(cjpeg-int sharedlib/cjpeg -dct int -outfile testoutint.jpg ${CMAKE_SOUR
 add_test(cjpeg-int-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgint.jpg testoutint.jpg)
 add_test(cjpeg-fast sharedlib/cjpeg -dct fast -opt -outfile testoutfst.jpg ${CMAKE_SOURCE_DIR}/testorig.ppm)
 add_test(cjpeg-fast-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgfst.jpg testoutfst.jpg)
+add_test(cjpeg-fast-100 sharedlib/cjpeg -dct fast -quality 100 -opt -outfile testoutfst100.jpg ${CMAKE_SOURCE_DIR}/testorig.ppm)
+add_test(cjpeg-fast-100-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgfst100.jpg testoutfst100.jpg)
 add_test(cjpeg-float sharedlib/cjpeg -dct float -outfile testoutflt.jpg ${CMAKE_SOURCE_DIR}/testorig.ppm)
 if(WITH_SIMD)
 add_test(cjpeg-float-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgflt.jpg testoutflt.jpg)
@@ -257,6 +257,8 @@ add_test(cjpeg-static-int cjpeg-static -dct int -outfile testoutint.jpg ${CMAKE_
 add_test(cjpeg-static-int-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgint.jpg testoutint.jpg)
 add_test(cjpeg-static-fast cjpeg-static -dct fast -opt -outfile testoutfst.jpg ${CMAKE_SOURCE_DIR}/testorig.ppm)
 add_test(cjpeg-static-fast-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgfst.jpg testoutfst.jpg)
+add_test(cjpeg-static-fast-100 cjpeg-static -dct fast -quality 100 -opt -outfile testoutfst100.jpg ${CMAKE_SOURCE_DIR}/testorig.ppm)
+add_test(cjpeg-static-fast-100-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgfst100.jpg testoutfst100.jpg)
 add_test(cjpeg-static-float cjpeg-static -dct float -outfile testoutflt.jpg ${CMAKE_SOURCE_DIR}/testorig.ppm)
 if(WITH_SIMD)
 add_test(cjpeg-static-float-cmp ${CMAKE_COMMAND} -E compare_files ${CMAKE_SOURCE_DIR}/testimgflt.jpg testoutflt.jpg)
@@ -328,7 +330,7 @@ add_custom_target(installer
  DEPENDS jpeg jpeg-static turbojpeg turbojpeg-static
  SOURCES libjpeg-turbo.nsi)

-install(TARGETS jpeg-static turbojpeg turbojpeg-static rdjpgcom wrjpgcom
+install(TARGETS jpeg-static turbojpeg turbojpeg-static rdjpgcom wrjpgcom jpgtest
  ARCHIVE DESTINATION lib
  LIBRARY DESTINATION lib
  RUNTIME DESTINATION bin
--- a/ChangeLog.txt
+++ b/ChangeLog.txt
@@ -1,14 +1,48 @@
-Significant changes since 1.0.1
-===============================
+1.1.0
+=====

-[1] Added emulation of the libjpeg v7 and v8b APIs and ABIs.  See
+[1] The algorithm used by the SIMD quantization function cannot produce correct
+results when the JPEG quality is >= 98 and the fast integer forward DCT is
+used.  Thus, the non-SIMD quantization function is now used for those cases,
+and libjpeg-turbo should now produce identical output to libjpeg v6b in all
+cases.
+
+[2] Despite the above, the fast integer forward DCT still degrades somewhat for
+JPEG qualities greater than 95, so TurboJPEG/OSS will now automatically use the
+slow integer forward DCT when generating JPEG images of quality 96 or greater.
+This reduces compression performance by as much as 15% for these high-quality
+images but is necessary to ensure that the images are perceptually lossless.
+It also ensures that the library can avoid the performance pitfall created by
+[1].
+
+[3] Ported jpgtest.cxx to pure C to avoid the need for a C++ compiler.
+
+[4] Fixed visual artifacts in grayscale JPEG compression caused by a typo in
+the RGB-to-chrominance lookup tables.
+
+[5] The Windows distribution packages now include the libjpeg run-time programs
+(cjpeg, etc.)
+
+[6] All packages now include jpgtest.
+
+[7] The TurboJPEG dynamic library now uses versioned symbols.
+
+[8] Added two new TurboJPEG API functions, tjEncodeYUV() and
+tjDecompressToYUV(), to replace the somewhat hackish TJ_YUV flag.
+
+
+1.0.90 (1.1 beta1)
+==================
+
+[1] Added emulation of the libjpeg v7 and v8 APIs and ABIs.  See
 README-turbo.txt for more details.  This feature was sponsored by CamTrace SAS.

 [2] Created a new CMake-based build system for the Visual C++ and MinGW builds.

 [3] TurboJPEG/OSS can now compress from/decompress to grayscale bitmaps.

-[4] jpgtest can now be used to test decompression performance only.
+[4] jpgtest can now be used to test decompression performance with existing
+JPEG images.

 [5] If the default install prefix (/opt/libjpeg-turbo) is used, then
 'make install' now creates /opt/libjpeg-turbo/lib32 and
@@ -18,7 +52,7 @@ packages.
 [6] All symbols in the libjpeg-turbo dynamic library are now versioned, even
 when the library is built with libjpeg v6b emulation.

-[7] Added arithmetic encoding and decoding support (can be disabled via
+[7] Added arithmetic encoding and decoding support (can be disabled with
 configure or CMake options)

 [8] Added a TJ_YUV flag to TurboJPEG/OSS which causes both the compressor and
@@ -30,8 +64,8 @@ allows the caller to determine the type of subsampling used in a JPEG image.
 [10] Added further protections against invalid Huffman codes.


-Significant changes since 1.0.0
-===============================
+1.0.1
+=====

 [1] The Huffman decoder will now handle erroneous Huffman codes (for instance,
 from a corrupt JPEG image.)  Previously, these would cause libjpeg-turbo to
@@ -44,8 +78,8 @@ be used instead of 4:2:0 when decompressing JPEG images using SSE2 code.
 INCOMPLETE_TYPES_BROKEN macro should be defined.


-Significant changes since 0.0.93
-================================
+1.0.0
+=====

 [1] 2983700: Further FreeBSD build tweaks (no longer necessary to specify
 --host when configuring on a 64-bit system)
@@ -69,18 +103,24 @@ contains just the 32-bit libjpeg-turbo libraries.
 unit tests now work on those architectures.


-Significant changes since 0.0.91
-================================
+0.0.93
+======

 [1] 2982659, Fixed x86-64 build on FreeBSD systems

 [2] 2988188: Added support for Windows 64-bit systems


-Significant changes since 0.0.90
-================================
+0.0.91
+======

 [1] Added documentation to .deb packages

 [2] 2968313: Fixed data corruption issues when decompressing large JPEG images
 and/or using buffered I/O with the libjpeg-turbo decompressor
+
+
+0.0.90
+======
+
+Initial release
--- a/Makefile.am
+++ b/Makefile.am
@@ -37,14 +37,9 @@ endif
 libturbojpeg_la_SOURCES = $(libjpeg_la_SOURCES) turbojpegl.c turbojpeg.h \
 	turbojpeg-mapfile

-if ANON_VERSION_SCRIPT
-
-libturbojpeg_la_LDFLAGS += $(ANON_VERSION_SCRIPT_FLAG)$(srcdir)/turbojpeg-mapfile
-
-endif
-
 if VERSION_SCRIPT

+libturbojpeg_la_LDFLAGS += $(VERSION_SCRIPT_FLAG)$(srcdir)/turbojpeg-mapfile
 libjpeg_la_LDFLAGS += $(VERSION_SCRIPT_FLAG)libjpeg.map

 endif
@@ -63,12 +58,12 @@ endif

 TSTHDRS = rrutil.h rrtimer.h

-bin_PROGRAMS = cjpeg djpeg jpegtran rdjpgcom wrjpgcom
-noinst_PROGRAMS = jpgtest jpegut
+bin_PROGRAMS = cjpeg djpeg jpegtran rdjpgcom wrjpgcom jpgtest
+noinst_PROGRAMS = jpegut

-jpgtest_SOURCES = $(TSTHDRS) jpgtest.cxx bmp.h bmp.c
+jpgtest_SOURCES = $(TSTHDRS) jpgtest.c bmp.h bmp.c

-jpgtest_LDADD = libturbojpeg.la
+jpgtest_LDADD = libturbojpeg.la -lm

 jpegut_SOURCES = $(TSTHDRS) jpegut.c bmp.h bmp.c

@@ -113,7 +108,7 @@ DOCS= README install.txt usage.txt wizard.txt example.c libjpeg.txt \
 TESTFILES= testorig.jpg testorig.ppm testimg.bmp testimgflt.jpg \
 	testimgfst.jpg testimgint.jpg testimgp.jpg testimgflt.ppm testimgfst.ppm \
 	testimgint.ppm testimgflt-nosimd.jpg testimgcrop.jpg testimgari.jpg \
-	testimgari.ppm
+	testimgari.ppm testimgfst100.jpg

 EXTRA_DIST = win release $(DOCS) $(TESTFILES) CMakeLists.txt \
 	sharedlib/CMakeLists.txt cmakescripts libjpeg.map.in
@@ -129,6 +124,8 @@ test: testclean all
 	cmp $(srcdir)/testimgint.jpg testoutint.jpg
 	./cjpeg -dct fast -opt -outfile testoutfst.jpg $(srcdir)/testorig.ppm
 	cmp $(srcdir)/testimgfst.jpg testoutfst.jpg
+	./cjpeg -dct fast -quality 100 -opt -outfile testoutfst100.jpg $(srcdir)/testorig.ppm
+	cmp $(srcdir)/testimgfst100.jpg testoutfst100.jpg
 	./cjpeg -dct float -outfile testoutflt.jpg $(srcdir)/testorig.ppm
 if WITH_SIMD
 	cmp $(srcdir)/testimgflt.jpg testoutflt.jpg
--- a/README-turbo.txt
+++ b/README-turbo.txt
@@ -2,22 +2,23 @@
 **     Background
 *******************************************************************************

-libjpeg-turbo is a high-speed version of libjpeg for x86 and x86-64 processors
-which uses SIMD instructions (MMX, SSE2, etc.) to accelerate baseline JPEG
-compression and decompression.  libjpeg-turbo is generally 2-4x as fast
+libjpeg-turbo is a derivative of libjpeg which uses SIMD instructions (MMX,
+SSE2, etc.) to accelerate baseline JPEG compression and decompression on x86
+and x86-64 systems.  On such systems, libjpeg-turbo is generally 2-4x as fast
 as the unmodified version of libjpeg, all else being equal.

 libjpeg-turbo was originally based on libjpeg/SIMD by Miyasaka Masaru, but
-the TigerVNC and VirtualGL projects made numerous enhancements to the codec,
-including improved support for Mac OS X, 64-bit support, support for 32-bit
-and big endian pixel formats, accelerated Huffman encoding/decoding, and
-various bug fixes.  The goal was to produce a fully open source codec that
-could replace the partially closed source TurboJPEG/IPP codec used by VirtualGL
-and TurboVNC.  libjpeg-turbo generally performs in the range of 80-120% of
-TurboJPEG/IPP.  It is faster in some areas but slower in others.
+the TigerVNC and VirtualGL projects made numerous enhancements to the codec in
+2009, including improved support for Mac OS X, 64-bit support, support for
+32-bit and big endian pixel formats (RGBX, XBGR, etc.), accelerated Huffman
+encoding/decoding, and various bug fixes.  The goal was to produce a fully open
+source codec that could replace the partially closed source TurboJPEG/IPP codec
+used by VirtualGL and TurboVNC.  libjpeg-turbo generally performs in the range
+of 80-120% of TurboJPEG/IPP.  It is faster in some areas but slower in others.

-It was decided to split libjpeg-turbo into a separate SDK so that other
-projects could take advantage of this technology.  The libjpeg-turbo shared
+In early 2010, libjpeg-turbo spun off into its own independent project, with
+the goal of making high-speed JPEG compression/decompression technology
+available to a broader range of users and developers.  The libjpeg-turbo shared
 libraries can be used as drop-in replacements for libjpeg on most systems.


@@ -25,21 +26,21 @@ libraries can be used as drop-in replacements for libjpeg on most systems.
 **     License
 *******************************************************************************

-Some of the optimizations to the Huffman encoder (jchuff.c) and decoder
-(jdhuff.c) were borrowed from VirtualGL, and thus any distribution of
-libjpeg-turbo which includes those optimizations must, as a whole, be subject
-to the terms of the wxWindows Library Licence, Version 3.1.  A copy of this
-license can be found in this directory under LICENSE.txt.  The wxWindows
-Library License is based on the LGPL but includes provisions which allow the
-Library to be statically linked into proprietary libraries and applications
-without requiring the resulting binaries to be distributed under the terms of
-the LGPL.
+The TurboJPEG/OSS wrapper, as well as some of the optimizations to the Huffman
+encoder (jchuff.c) and decoder (jdhuff.c), were borrowed from VirtualGL, and
+thus any distribution of libjpeg-turbo which includes those files must, as a
+whole, be subject to the terms of the wxWindows Library Licence, Version 3.1.
+A copy of this license can be found in this directory under LICENSE.txt.  The
+wxWindows Library License is based on the LGPL but includes provisions which
+allow the Library to be statically linked into proprietary libraries and
+applications without requiring the resulting binaries to be distributed under
+the terms of the LGPL.

-The rest of the source code, apart from the Huffman codec optimizations, falls
-under a less restrictive, BSD-style license (see README.)  You can choose to
-distribute libjpeg-turbo, as a whole, under this BSD-style license by simply
-replacing the optimized jchuff.c and jdhuff.c with their unoptimized
-counterparts from the libjpeg v6b source.
+The rest of the source code, apart from TurboJPEG/OSS and the Huffman codec
+optimizations, falls under a less restrictive, BSD-style license (see README.)
+You can choose to distribute libjpeg-turbo, as a whole, under this BSD-style
+license by simply removing TurboJPEG/OSS and replacing the optimized jchuff.c
+and jdhuff.c with their unoptimized counterparts from the libjpeg v6b source.


 *******************************************************************************
@@ -51,7 +52,7 @@ Replacing libjpeg at Run Time
 =============================

 If a Unix application is dynamically linked with libjpeg, then you can replace
-libjpeg with libjpeg-turbo at run time by manipulating the LD_LIBRARY_PATH.
+libjpeg with libjpeg-turbo at run time by manipulating LD_LIBRARY_PATH.
 For instance:

  [Using libjpeg]
@@ -75,28 +76,29 @@ links to the libjpeg dynamic library located in /opt/libjpeg-turbo/{lib}.  This
 will effectively accelerate every dynamically linked libjpeg application on the
 system.

-The Windows distribution of the libjpeg-turbo SDK installs the libjpeg-turbo
-DLL (jpeg62.dll, jpeg7.dll, or jpeg8.dll, depending on whether libjpeg v6b,
-v7, or v8 emulation is enabled) into c:\libjpeg-turbo[64]\bin, and the PATH
+The libjpeg-turbo SDK for Visual C++ installs the libjpeg-turbo DLL
+(jpeg62.dll, jpeg7.dll, or jpeg8.dll, depending on whether libjpeg v6b, v7, or
+v8 emulation is enabled) into c:\libjpeg-turbo[64]\bin, and the PATH
 environment variable can be modified such that this directory is searched
 before any others that might contain a libjpeg DLL.  However, if a libjpeg
 DLL exists in an application's install directory, then Windows will load this
 DLL first whenever the application is launched.  Thus, if an application ships
 with jpeg62.dll, jpeg7.dll, or jpeg8.dll, then back up the application's
-version of this DLL and copy c:\libjpeg-turbo\bin\jpeg*.dll into the
+version of this DLL and copy c:\libjpeg-turbo[64]\bin\jpeg*.dll into the
 application's install directory to accelerate it.

-The version of the libjpeg-turbo DLL distributed with the "official"
-libjpeg-turbo SDK requires the Visual C++ 2008 C run time DLL (msvcr90.dll).
+The version of the libjpeg-turbo DLL distributed in the libjpeg-turbo SDK for
+Visual C++ requires the Visual C++ 2008 C run time DLL (msvcr90.dll).
 msvcr90.dll ships with more recent versions of Windows, but users of older
 Windows releases can obtain it from the Visual C++ 2008 Redistributable
 Package, which is available as a free download from Microsoft's web site.

 NOTE:  Features of libjpeg which require passing a C run time structure, such
 as a file handle, from an application to libjpeg will probably not work with
-the "official" version of the libjpeg-turbo DLL unless the application is also
-built to use the Visual C++ 2008 C run time DLL.  In particular, this affects
-jpeg_stdio_dest() and jpeg_stdio_src().
+the version of the libjpeg-turbo DLL distributed in the libjpeg-turbo SDK for
+Visual C++, unless the application is also built to use the Visual C++ 2008 C
+run time DLL.  In particular, this affects jpeg_stdio_dest() and
+jpeg_stdio_src().

 Mac applications typically embed their own copies of the libjpeg dylib inside
 the (hidden) application bundle, so it is not possible to globally replace
@@ -117,8 +119,8 @@ VirtualGL 2.1.x and TurboVNC 0.6 (and prior.)  libjpeg-turbo contains a wrapper
 library (TurboJPEG/OSS) that emulates the TurboJPEG API using libjpeg-turbo
 instead of the closed source Intel Performance Primitives.  You can replace the
 TurboJPEG/IPP package on Linux systems with the libjpeg-turbo package in order
-to make existing releases of VirtualGL 2.1.x and TurboVNC use the new codec at
-run time.  Note that the 64-bit libjpeg-turbo packages contain only 64-bit
+to make existing releases of VirtualGL 2.1.x and TurboVNC 0.x use the new codec
+at run time.  Note that the 64-bit libjpeg-turbo packages contain only 64-bit
 binaries, whereas the TurboJPEG/IPP 64-bit packages contained both 64-bit and
 32-bit binaries.  Thus, to replace a TurboJPEG/IPP 64-bit package, install
 both the 64-bit and 32-bit versions of libjpeg-turbo.
@@ -152,11 +154,11 @@ If using MinGW, then set

 Building against libjpeg-turbo is useful, for instance, if you want to build an
 application that leverages the libjpeg-turbo colorspace extensions (see below.)
-On Linux and Solaris systems, you would still need to manipulate the
-LD_LIBRARY_PATH or sym links appropriately to use libjpeg-turbo at run time.
-On such systems, you can pass -R /opt/libjpeg-turbo/{lib} to the linker to
-force the use of libjpeg-turbo at run time rather than libjpeg (also useful if
-you want to leverage the colorspace extensions), or you can link against the
+On Linux and Solaris systems, you would still need to manipulate
+LD_LIBRARY_PATH or create appropriate sym links to use libjpeg-turbo at run
+time.  On such systems, you can pass -R /opt/libjpeg-turbo/{lib} to the linker
+to force the use of libjpeg-turbo at run time rather than libjpeg (also useful
+if you want to leverage the colorspace extensions), or you can link against the
 libjpeg-turbo static library.

 To force a Linux, Solaris, or MinGW application to link against the static
@@ -168,8 +170,8 @@ On OS X, simply add /opt/libjpeg-turbo/lib/libjpeg.a to the linker command
 line (this also works on Linux and Solaris.)

 To build Visual C++ applications using libjpeg-turbo, add
-c:\libjpeg-turbo[64]\include to your system or user INCLUDE environment
-variable and c:\libjpeg-turbo[64]\lib to your system or user LIB environment
+c:\libjpeg-turbo[64]\include to the system or user INCLUDE environment
+variable and c:\libjpeg-turbo[64]\lib to the system or user LIB environment
 variable, and then link against either jpeg.lib (to use the DLL version of
 libjpeg-turbo) or jpeg-static.lib (to use the static version of libjpeg-turbo.)

@@ -227,7 +229,7 @@ releases will also be backward-incompatible.

 By passing an argument of --with-jpeg7 or --with-jpeg8 to configure, or an
 argument of -DWITH_JPEG7=1 or -DWITH_JPEG8=1 to cmake, you can build a version
-of libjpeg-turbo which emulates the libjpeg v7 or v8b API/ABI, so that programs
+of libjpeg-turbo which emulates the libjpeg v7 or v8 API/ABI, so that programs
 which are built against libjpeg v7 or v8 can be run with libjpeg-turbo.  The
 following section describes which libjpeg v7+ features are supported and which
 aren't.
@@ -272,3 +274,31 @@ Not supported:

 -- jpegtran: Scaling
   Seems to depend on the DCT scaling feature, which isn't supported.
+
+
+*******************************************************************************
+**     Performance pitfalls
+*******************************************************************************
+
+===============
+Restart Markers
+===============
+
+The optimized Huffman decoder in libjpeg-turbo does not handle restart markers
+in a way that makes libjpeg happy, so it is necessary to use the slow Huffman
+decoder when decompressing a JPEG image that has restart markers.  This can
+cause the decompression performance to drop by as much as 20%, but the
+performance will still be much much greater than that of libjpeg v6b.  Many
+consumer packages, such as PhotoShop, use restart markers when generating JPEG
+images, so images generated by those programs will experience this issue.
+
+===============================================
+Fast Integer Forward DCT at High Quality Levels
+===============================================
+
+The algorithm used by the SIMD-accelerated quantization function cannot produce
+correct results whenever the fast integer forward DCT is used along with a JPEG
+quality of 98-100.  Thus, libjpeg-turbo must use the non-SIMD quantization
+function in those cases.  This causes performance to drop by as much as 40%.
+It is therefore strongly advised that you use the slow integer forward DCT
+whenever encoding images with a JPEG quality of 98 or higher.
--- a/bmp.h
+++ b/bmp.h
@@ -1,5 +1,6 @@
 /* Copyright (C)2004 Landmark Graphics Corporation
 * Copyright (C)2005 Sun Microsystems, Inc.
+ * Copyright (C)2011 D. R. Commander
 *
 * This library is free software and may be redistributed and/or modified under
 * the terms of the wxWindows Library License, Version 3.1 or (at your option)
@@ -19,7 +20,7 @@
 #define __BMP_H__

 #define BMPPIXELFORMATS 6
-enum BMPPIXELFORMAT {BMP_RGB=0, BMP_RGBA, BMP_BGR, BMP_BGRA, BMP_ABGR, BMP_ARGB};
+enum BMPPIXELFORMAT {BMP_RGB=0, BMP_RGBX, BMP_BGR, BMP_BGRX, BMP_XBGR, BMP_XRGB};

 #ifdef __cplusplus
 extern "C" {
--- a/configure.ac
+++ b/configure.ac
@@ -2,7 +2,7 @@
 # Process this file with autoconf to produce a configure script.

 AC_PREREQ([2.56])
-AC_INIT([libjpeg-turbo], [1.0.90])
+AC_INIT([libjpeg-turbo], [1.1.0])
 BUILD=`date +%Y%m%d`

 AM_INIT_AUTOMAKE([-Wall foreign dist-bzip2])
@@ -13,10 +13,9 @@ AC_DEFINE([HAVE_PROTOTYPES], 1, [Define if your compiler supports prototypes])

 # Checks for programs.
 SAVED_CFLAGS=${CFLAGS}
-SAVED_CXXFLAGS=${CXXFLAGS}
+SAVED_CPPFLAGS=${CPPFLAGS}
 AC_PROG_CPP
 AC_PROG_CC
-AC_PROG_CXX
 AC_PROG_INSTALL
 AC_PROG_LIBTOOL
 AC_PROG_LN_S
@@ -32,8 +31,8 @@ if test "x${GCC}" = "xyes"; then
  if test "x${SAVED_CFLAGS}" = "x"; then
    CFLAGS=-O3
  fi
-  if test "x${SAVED_CXXFLAGS}" = "x"; then
-    CXXFLAGS=-O3
+  if test "x${SAVED_CPPFLAGS}" = "x"; then
+    CPPFLAGS=-Wall
  fi
 fi

@@ -42,9 +41,6 @@ if test "x${SUNCC}" = "xyes"; then
  if test "x${SAVED_CFLAGS}" = "x"; then
    CFLAGS=-xO5
  fi
-  if test "x${SAVED_CXXFLAGS}" = "x"; then
-    CXXFLAGS=-xO5
-  fi
 fi

 # Checks for libraries.
@@ -116,7 +112,7 @@ if test "x$JPEG_LIB_VERSION" = "x"; then
    AC_ARG_WITH([jpeg7],
        AC_HELP_STRING([--with-jpeg7], [Emulate libjpeg v7 API/ABI (this makes libjpeg-turbo backward incompatible with libjpeg v6b.)]))
    AC_ARG_WITH([jpeg8],
-        AC_HELP_STRING([--with-jpeg8], [Emulate libjpeg v8b API/ABI (this makes libjpeg-turbo backward incompatible with libjpeg v6b.)]))
+        AC_HELP_STRING([--with-jpeg8], [Emulate libjpeg v8 API/ABI (this makes libjpeg-turbo backward incompatible with libjpeg v6b.)]))
    if test "x${with_jpeg8}" = "xyes"; then
        JPEG_LIB_VERSION=80
    else
@@ -176,36 +172,13 @@ if test "x$VERSION_SCRIPT_FLAG" = "x"; then
  VERSION_SCRIPT=no
  AC_MSG_RESULT(no)
 fi
-
-ANON_VERSION_SCRIPT=yes
-AC_MSG_CHECKING([whether the linker supports anonymous version scripts])
-LDFLAGS="$SAVED_LDFLAGS -Wl,--version-script,conftest.map"
-cat > conftest.map <<EOF
-{
-  global: main;
-  local: *;
-};
-EOF
-AC_LINK_IFELSE(AC_LANG_PROGRAM([], []),
-  [ANON_VERSION_SCRIPT_FLAG=-Wl,--version-script,; AC_MSG_RESULT([yes (GNU style)])], [])
-if test "x$ANON_VERSION_SCRIPT_FLAG" = "x"; then
-  LDFLAGS="$SAVED_LDFLAGS -Wl,-M,conftest.map"
-  AC_LINK_IFELSE(AC_LANG_PROGRAM([], []),
-    [ANON_VERSION_SCRIPT_FLAG=-Wl,-M,; AC_MSG_RESULT([yes (Sun style)])], [])
-fi
-if test "x$ANON_VERSION_SCRIPT_FLAG" = "x"; then
-  ANON_VERSION_SCRIPT=no
-  AC_MSG_RESULT(no)
-fi
 LDFLAGS="$SAVED_LDFLAGS"

-AC_MSG_CHECKING([whether to use version script when building libjpeg])
+AC_MSG_CHECKING([whether to use version script when building libjpeg-turbo])
 AC_MSG_RESULT($VERSION_SCRIPT)

 AM_CONDITIONAL(VERSION_SCRIPT, test "x$VERSION_SCRIPT" = "xyes")
-AM_CONDITIONAL(ANON_VERSION_SCRIPT, test "x$ANON_VERSION_SCRIPT" = "xyes")
 AC_SUBST(VERSION_SCRIPT_FLAG)
-AC_SUBST(ANON_VERSION_SCRIPT_FLAG)

 AC_MSG_CHECKING([whether to include arithmetic encoding support])
 AC_ARG_WITH([arith-enc],
--- a/jccolor.c
+++ b/jccolor.c
@@ -122,7 +122,7 @@ static const unsigned char green_lut[256] = {
  113, 113, 114, 114, 115, 116, 116, 117, 117, 118, 119, 119,
  120, 120, 121, 122, 122, 123, 123, 124, 124, 125, 126, 126,
  127, 127, 128, 129, 129, 130, 130, 131, 131, 132, 133, 133,
-  34, 134, 135, 136, 136, 137, 137, 138, 139, 139, 140, 140,
+  134, 134, 135, 136, 136, 137, 137, 138, 139, 139, 140, 140,
  141, 141, 142, 143, 143, 144, 144, 145, 146, 146, 147, 147,
  148, 149, 149, 150
 };
@@ -258,14 +258,12 @@ rgb_gray_convert (j_compress_ptr cinfo,
 		  JSAMPARRAY input_buf, JSAMPIMAGE output_buf,
 		  JDIMENSION output_row, int num_rows)
 {
-  my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert;
  #if BITS_IN_JSAMPLE != 8
  register INT32 * ctab = cconvert->rgb_ycc_tab;
  #endif
  register JSAMPROW inptr;
  register JSAMPROW outptr;
  JSAMPLE *maxoutptr;
-  register JDIMENSION col;
  JDIMENSION num_cols = cinfo->image_width;
  int rindex = rgb_red[cinfo->in_color_space];
  int gindex = rgb_green[cinfo->in_color_space];
--- a/jcdctmgr.c
+++ b/jcdctmgr.c
@@ -4,6 +4,7 @@
 * Copyright (C) 1994-1996, Thomas G. Lane.
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
 * Copyright 2009 Pierre Ossman <ossman@cendio.se> for Cendio AB
+ * Copyright (C) 2011 D. R. Commander
 * This file is part of the Independent JPEG Group's software.
 * For conditions of distribution and use, see the accompanying README file.
 *
@@ -39,6 +40,8 @@ typedef JMETHOD(void, float_quantize_method_ptr,
                (JCOEFPTR coef_block, FAST_FLOAT * divisors,
                 FAST_FLOAT * workspace));

+METHODDEF(void) quantize (JCOEFPTR, DCTELEM *, DCTELEM *);
+
 typedef struct {
  struct jpeg_forward_dct pub;	/* public fields */

@@ -160,7 +163,7 @@ flss (UINT16 val)
 * of in a consecutive manner, yet again in order to allow SIMD
 * routines.
 */
-LOCAL(void)
+LOCAL(int)
 compute_reciprocal (UINT16 divisor, DCTELEM * dtbl)
 {
  UDCTELEM2 fq, fr;
@@ -189,6 +192,9 @@ compute_reciprocal (UINT16 divisor, DCTELEM * dtbl)
  dtbl[DCTSIZE2 * 1] = (DCTELEM) c;       /* correction + roundfactor */
  dtbl[DCTSIZE2 * 2] = (DCTELEM) (1 << (sizeof(DCTELEM)*8*2 - r));  /* scale */
  dtbl[DCTSIZE2 * 3] = (DCTELEM) r - sizeof(DCTELEM)*8; /* shift */
+
+  if(r <= 16) return 0;
+  else return 1;
 }

 /*
@@ -232,7 +238,9 @@ start_pass_fdctmgr (j_compress_ptr cinfo)
      }
      dtbl = fdct->divisors[qtblno];
      for (i = 0; i < DCTSIZE2; i++) {
-	compute_reciprocal(qtbl->quantval[i] << 3, &dtbl[i]);
+	if(!compute_reciprocal(qtbl->quantval[i] << 3, &dtbl[i])
+	  && fdct->quantize == jsimd_quantize)
+	  fdct->quantize = quantize;
      }
      break;
 #endif
@@ -266,10 +274,12 @@ start_pass_fdctmgr (j_compress_ptr cinfo)
 	}
 	dtbl = fdct->divisors[qtblno];
 	for (i = 0; i < DCTSIZE2; i++) {
-	  compute_reciprocal(
+	  if(!compute_reciprocal(
 	    DESCALE(MULTIPLY16V16((INT32) qtbl->quantval[i],
 				  (INT32) aanscales[i]),
-		    CONST_BITS-3), &dtbl[i]);
+		    CONST_BITS-3), &dtbl[i])
+	    && fdct->quantize == jsimd_quantize)
+	    fdct->quantize = quantize;
 	}
      }
      break;
--- a/jdcolor.c
+++ b/jdcolor.c
@@ -223,7 +223,6 @@ gray_rgb_convert (j_decompress_ptr cinfo,
 {
  register JSAMPROW inptr, outptr;
  JSAMPLE *maxinptr;
-  register JDIMENSION col;
  JDIMENSION num_cols = cinfo->output_width;
  int rindex = rgb_red[cinfo->out_color_space];
  int gindex = rgb_green[cinfo->out_color_space];
--- a/jpegut.c
+++ b/jpegut.c
@@ -1,6 +1,6 @@
 /* Copyright (C)2004 Landmark Graphics Corporation
 * Copyright (C)2005 Sun Microsystems, Inc.
- * Copyright (C)2009-2010 D. R. Commander
+ * Copyright (C)2009-2011 D. R. Commander
 *
 * This library is free software and may be redistributed and/or modified under
 * the terms of the wxWindows Library License, Version 3.1 or (at your option)
@@ -107,12 +107,13 @@ void dumpbuf(unsigned char *buf, int w, int h, int ps, int flags)
 {
 	int roffset=(flags&TJ_BGR)?2:0, goffset=1, boffset=(flags&TJ_BGR)?0:2, i,
 		j;
+	printf("\n");
 	for(i=0; i<h; i++)
 	{
 		for(j=0; j<w; j++)
 		{
 			printf("%.3d/%.3d/%.3d ", buf[(w*i+j)*ps+roffset],
-				buf[(w*i+j)*ps+roffset], buf[(w*i+j)*ps+roffset]);
+				buf[(w*i+j)*ps+goffset], buf[(w*i+j)*ps+boffset]);
 		}
 		printf("\n");
 	}
@@ -225,8 +226,7 @@ int checkbuf(unsigned char *buf, int w, int h, int ps, int subsamp, int flags)

 #define PAD(v, p) ((v+(p)-1)&(~((p)-1)))

-int checkbufyuv(unsigned char *buf, unsigned long size, int w, int h,
-	int subsamp)
+int checkbufyuv(unsigned char *buf, int w, int h, int subsamp)
 {
 	int i, j;
 	int hsf=_hsf[subsamp], vsf=_vsf[subsamp];
@@ -234,13 +234,6 @@ int checkbufyuv(unsigned char *buf, unsigned long size, int w, int h,
 	int cw=pw/hsf, ch=ph/vsf;
 	int ypitch=PAD(pw, 4), uvpitch=PAD(cw, 4);
 	int retval=1;
-	unsigned long correctsize=ypitch*ph + (subsamp==TJ_GRAYSCALE? 0:uvpitch*ch*2);
-
-	if(size!=correctsize)
-	{
-		printf("\nIncorrect size %lu.  Should be %lu\n", size, correctsize);
-		retval=0;  goto bailout;
-	}

 	for(i=0; i<16; i++)
 	{
@@ -349,17 +342,15 @@ void gentestjpeg(tjhandle hnd, unsigned char *jpegbuf, unsigned long *size,
 	char tempstr[1024];  unsigned char *bmpbuf=NULL;
 	const char *pixformat;  double t;

-	if(yuv==YUVENCODE) flags|=TJ_YUV;
-
 	if(flags&TJ_BGR)
 	{
 		if(ps==3) pixformat="BGR";
-		else {if(flags&TJ_ALPHAFIRST) pixformat="ABGR";  else pixformat="BGRA";}
+		else {if(flags&TJ_ALPHAFIRST) pixformat="XBGR";  else pixformat="BGRX";}
 	}
 	else
 	{
 		if(ps==3) pixformat="RGB";
-		else {if(flags&TJ_ALPHAFIRST) pixformat="ARGB";  else pixformat="RGBA";}
+		else {if(flags&TJ_ALPHAFIRST) pixformat="XRGB";  else pixformat="RGBX";}
 	}
 	if(ps==1) pixformat="Grayscale";
 	if(yuv==YUVENCODE)
@@ -374,10 +365,20 @@ void gentestjpeg(tjhandle hnd, unsigned char *jpegbuf, unsigned long *size,
 		printf("ERROR: Could not allocate buffer\n");  bailout();
 	}
 	initbuf(bmpbuf, w, h, ps, flags);
-	memset(jpegbuf, 0, TJBUFSIZE(w, h));
+	memset(jpegbuf, 0,
+		yuv==YUVENCODE? TJBUFSIZEYUV(w, h, subsamp):TJBUFSIZE(w, h));

 	t=rrtime();
-	_catch(tjCompress(hnd, bmpbuf, w, 0, h, ps, jpegbuf, size, subsamp, qual, flags));
+	if(yuv==YUVENCODE)
+	{
+		_catch(tjEncodeYUV(hnd, bmpbuf, w, 0, h, ps, jpegbuf, subsamp, flags));
+		*size=TJBUFSIZEYUV(w, h, subsamp);
+	}
+	else
+	{
+		_catch(tjCompress(hnd, bmpbuf, w, 0, h, ps, jpegbuf, size, subsamp, qual,
+			flags));
+	}
 	t=rrtime()-t;

 	if(yuv==YUVENCODE)
@@ -389,8 +390,8 @@ void gentestjpeg(tjhandle hnd, unsigned char *jpegbuf, unsigned long *size,
 	writejpeg(jpegbuf, *size, tempstr);
 	if(yuv==YUVENCODE)
 	{
-		if(checkbufyuv(jpegbuf, *size, w, h, subsamp)) printf("Passed.");
-		else printf("FAILED!");
+		if(checkbufyuv(jpegbuf, w, h, subsamp)) printf("Passed.");
+		else {printf("FAILED!");  exitstatus=-1;}
 	}
 	else printf("Done.");
 	printf("  %f ms\n  Result in %s\n", t*1000., tempstr);
@@ -405,23 +406,18 @@ void gentestbmp(tjhandle hnd, unsigned char *jpegbuf, unsigned long jpegsize,
 	unsigned char *bmpbuf=NULL;
 	const char *pixformat;  int _w=0, _h=0;  double t;
 	unsigned long size=0;
-	int hsf=_hsf[subsamp], vsf=_vsf[subsamp];
-	int pw=PAD(w, hsf), ph=PAD(h, vsf);
-	int cw=pw/hsf, ch=ph/vsf;
-	int ypitch=PAD(pw, 4), uvpitch=PAD(cw, 4);

-	if(yuv==YUVDECODE) flags|=TJ_YUV;
-	else if(yuv==YUVENCODE) return;
+	if(yuv==YUVENCODE) return;

 	if(flags&TJ_BGR)
 	{
 		if(ps==3) pixformat="BGR";
-		else {if(flags&TJ_ALPHAFIRST) pixformat="ABGR";  else pixformat="BGRA";}
+		else {if(flags&TJ_ALPHAFIRST) pixformat="XBGR";  else pixformat="BGRX";}
 	}
 	else
 	{
 		if(ps==3) pixformat="RGB";
-		else {if(flags&TJ_ALPHAFIRST) pixformat="ARGB";  else pixformat="RGBA";}
+		else {if(flags&TJ_ALPHAFIRST) pixformat="XRGB";  else pixformat="RGBX";}
 	}
 	if(ps==1) pixformat="Grayscale";
 	if(yuv==YUVDECODE)
@@ -436,30 +432,39 @@ void gentestbmp(tjhandle hnd, unsigned char *jpegbuf, unsigned long jpegsize,
 		printf("Incorrect JPEG header\n");  bailout();
 	}

-	if(yuv==YUVDECODE)
-		size=ypitch*ph + (subsamp==TJ_GRAYSCALE? 0:uvpitch*ch*2);
-	else
-		size=w*h*ps;
-	if((bmpbuf=(unsigned char *)malloc(size+1))==NULL)
+	if(yuv==YUVDECODE) size=TJBUFSIZEYUV(w, h, subsamp);
+	else size=w*h*ps+1;
+	if((bmpbuf=(unsigned char *)malloc(size))==NULL)
 	{
 		printf("ERROR: Could not allocate buffer\n");  bailout();
 	}
-	memset(bmpbuf, 0, size+1);
+	memset(bmpbuf, 0, size);

 	t=rrtime();
-	_catch(tjDecompress(hnd, jpegbuf, jpegsize, bmpbuf, w, w*ps, h, ps, flags));
+	if(yuv==YUVDECODE)
+	{
+		_catch(tjDecompressToYUV(hnd, jpegbuf, jpegsize, bmpbuf, flags));
+	}
+	else
+	{
+		_catch(tjDecompress(hnd, jpegbuf, jpegsize, bmpbuf, w, w*ps, h, ps,
+			flags));
+	}
 	t=rrtime()-t;

 	if(yuv==YUVDECODE)
 	{
-		if(checkbufyuv(bmpbuf, size, pw, ph, subsamp))
-			printf("Passed.");
-		else printf("FAILED!");
+		if(checkbufyuv(bmpbuf, w, h, subsamp)) printf("Passed.");
+		else {printf("FAILED!");  exitstatus=-1;}
 	}
 	else
 	{
 		if(checkbuf(bmpbuf, w, h, ps, subsamp, flags)) printf("Passed.");
-		else {printf("FAILED!");  dumpbuf(bmpbuf, w, h, ps, flags);}
+		else
+		{
+			printf("FAILED!");  exitstatus=-1;
+			dumpbuf(bmpbuf, w, h, ps, flags);
+		}
 	}
 	printf("  %f ms\n\n", t*1000.);

@@ -472,7 +477,8 @@ void dotest(int w, int h, int ps, int subsamp, char *basefilename)
 	tjhandle hnd=NULL, dhnd=NULL;  unsigned char *jpegbuf=NULL;
 	unsigned long size;

-	if((jpegbuf=(unsigned char *)malloc(TJBUFSIZE(w, h))) == NULL)
+	size=(yuv==YUVENCODE? TJBUFSIZEYUV(w, h, subsamp):TJBUFSIZE(w, h));
+	if((jpegbuf=(unsigned char *)malloc(size)) == NULL)
 	{
 		puts("ERROR: Could not allocate buffer.");  bailout();
 	}
@@ -542,9 +548,9 @@ void dotest1(void)
 			{
 				bmpbuf[i2*4]=pixels[i2%9][2];
 				bmpbuf[i2*4+1]=pixels[i2%9][1];
-				bmpbuf[i2*2+2]=pixels[i2%9][0];
+				bmpbuf[i2*4+2]=pixels[i2%9][0];
 			}
-			_catch(tjCompress(hnd, bmpbuf, i, i*4, j, 4,
+			_catch(tjCompress(hnd, bmpbuf, i, 0, j, 4,
 				jpgbuf, &size, TJ_444, 100, TJ_BGR));
 			free(bmpbuf);  bmpbuf=NULL;  free(jpgbuf);  jpgbuf=NULL;

@@ -553,12 +559,12 @@ void dotest1(void)
 			{
 				printf("Memory allocation failure\n");  bailout();
 			}
-			for(i2=0; i2<j*i*4; i2++)
+			for(i2=0; i2<j*i; i2++)
 			{
-				if(i2%2==0) bmpbuf[i2]=0xFF;
-				else bmpbuf[i2]=0;
+				if(i2%2==0) bmpbuf[i2*4]=bmpbuf[i2*4+1]=bmpbuf[i2*4+2]=0xFF;
+				else bmpbuf[i2*4]=bmpbuf[i2*4+1]=bmpbuf[i2*4+2]=0;
 			}
-			_catch(tjCompress(hnd, bmpbuf, j, j*4, i, 4,
+			_catch(tjCompress(hnd, bmpbuf, j, 0, i, 4,
 				jpgbuf, &size, TJ_444, 100, TJ_BGR));
 			free(bmpbuf);  bmpbuf=NULL;  free(jpgbuf);  jpgbuf=NULL;
 		}
@@ -593,6 +599,12 @@ int main(int argc, char *argv[])
 		yuv=YUVDECODE;
 		dotest(48, 48, 3, TJ_444, "test");
 		dotest(35, 39, 3, TJ_444, "test");
+		dotest(48, 48, 3, TJ_422, "test");
+		dotest(39, 41, 3, TJ_422, "test");
+		dotest(48, 48, 3, TJ_420, "test");
+		dotest(41, 35, 3, TJ_420, "test");
+		dotest(48, 48, 3, TJ_GRAYSCALE, "test");
+		dotest(35, 39, 3, TJ_GRAYSCALE, "test");
 		dotest(48, 48, 1, TJ_GRAYSCALE, "test");
 		dotest(39, 41, 1, TJ_GRAYSCALE, "test");
 	}
--- a/jpgtest.cxx
+++ b/jpgtest.cxx
@@ -1,6 +1,6 @@
 /* Copyright (C)2004 Landmark Graphics Corporation
 * Copyright (C)2005, 2006 Sun Microsystems, Inc.
- * Copyright (C)2009-2010 D. R. Commander
+ * Copyright (C)2009-2011 D. R. Commander
 *
 * This library is free software and may be redistributed and/or modified under
 * the terms of the wxWindows Library License, Version 3.1 or (at your option)
@@ -40,11 +40,9 @@ const int _flags[BMPPIXELFORMATS]={0, 0, TJ_BGR, TJ_BGR,
 const int _rindex[BMPPIXELFORMATS]={0, 0, 2, 2, 3, 1};
 const int _gindex[BMPPIXELFORMATS]={1, 1, 1, 1, 2, 2};
 const int _bindex[BMPPIXELFORMATS]={2, 2, 0, 0, 1, 3};
-const char *_pfname[]={"RGB", "RGBA", "BGR", "BGRA", "ABGR", "ARGB"};
+const char *_pfname[]={"RGB", "RGBX", "BGR", "BGRX", "XBGR", "XRGB"};
 const char *_subnamel[NUMSUBOPT]={"4:4:4", "4:2:2", "4:2:0", "GRAY"};
 const char *_subnames[NUMSUBOPT]={"444", "422", "420", "GRAY"};
-const int _hsf[NUMSUBOPT]={1, 2, 2, 1};
-const int _vsf[NUMSUBOPT]={1, 1, 2, 1};

 void printsigfig(double val, int figs)
 {
@@ -64,43 +62,39 @@ void printsigfig(double val, int figs)
 	printf(format, val);
 }

-void dotest(unsigned char *srcbuf, int w, int h, BMPPIXELFORMAT pf, int bu,
+void dotest(unsigned char *srcbuf, int w, int h, int pf, int bu,
 	int jpegsub, int qual, char *filename, int dotile, int useppm, int quiet)
 {
 	char tempstr[1024];
-	FILE *outfile=NULL;  tjhandle hnd;
+	FILE *outfile=NULL;  tjhandle hnd=NULL;
 	unsigned char **jpegbuf=NULL, *rgbbuf=NULL;
-	rrtimer timer; double elapsed;
-	int jpgbufsize=0, i, j, tilesizex, tilesizey, numtilesx, numtilesy, ITER;
+	double start, elapsed;
+	int jpgbufsize=0, i, j, tilesizex=w, tilesizey=h, numtilesx=1, numtilesy=1,
+		ITER;
 	unsigned long *comptilesize=NULL;
 	int flags=(forcemmx?TJ_FORCEMMX:0)|(forcesse?TJ_FORCESSE:0)
 		|(forcesse2?TJ_FORCESSE2:0)|(forcesse3?TJ_FORCESSE3:0)
 		|(fastupsample?TJ_FASTUPSAMPLE:0);
-	int ps=_ps[pf];
-	int pitch=w*ps, yuvsize;
-	int hsf=_hsf[jpegsub], vsf=_vsf[jpegsub];
-	int pw=PAD(w, hsf), ph=PAD(h, vsf);
-	int cw=pw/hsf, ch=ph/vsf;
-	int ypitch=PAD(pw, 4), uvpitch=PAD(cw, 4);
+	int ps=_ps[pf], tilen;
+	int pitch=w*ps, yuvsize=0;

 	flags |= _flags[pf];
 	if(bu) flags |= TJ_BOTTOMUP;
-	if(yuv==YUVENCODE) flags |= TJ_YUV;

-	yuvsize=ypitch*ph + (jpegsub==TJ_GRAYSCALE? 0:uvpitch*ch*2);
-	if((rgbbuf=(unsigned char *)malloc(max(yuvsize, pitch*h))) == NULL)
+	if(yuv==YUVENCODE) yuvsize=TJBUFSIZEYUV(w, h, jpegsub);
+	if((rgbbuf=(unsigned char *)malloc(max(yuvsize, pitch*h+1))) == NULL)
 		_throwunix("allocating image buffer");

 	if(!quiet)
 	{
 		if(yuv==YUVENCODE)
-			printf("\n>>>>>  %s (%s) <--> YUV %s  <<<<<\n", _pfname[pf],
+			printf(">>>>>  %s (%s) <--> YUV %s  <<<<<\n", _pfname[pf],
 				bu?"Bottom-up":"Top-down", _subnamel[jpegsub]);
 		else
-			printf("\n>>>>>  %s (%s) <--> JPEG %s Q%d  <<<<<\n", _pfname[pf],
+			printf(">>>>>  %s (%s) <--> JPEG %s Q%d  <<<<<\n", _pfname[pf],
 				bu?"Bottom-up":"Top-down", _subnamel[jpegsub], qual);
 	}
-	if(yuv==YUVDECODE) dotile=0;
+	if(yuv) dotile=0;
 	if(dotile) {tilesizex=tilesizey=4;}  else {tilesizex=w;  tilesizey=h;}

 	do
@@ -115,30 +109,46 @@ void dotest(unsigned char *srcbuf, int w, int h, BMPPIXELFORMAT pf, int bu,
 		memset(jpegbuf, 0, sizeof(unsigned char *)*numtilesx*numtilesy);
 		for(i=0; i<numtilesx*numtilesy; i++)
 		{
-			if((jpegbuf[i]=(unsigned char *)malloc(TJBUFSIZE(tilesizex, tilesizey))) == NULL)
+			if((jpegbuf[i]=(unsigned char *)malloc(
+				yuv==YUVENCODE? TJBUFSIZEYUV(tilesizex, tilesizey, jpegsub)
+					: TJBUFSIZE(tilesizex, tilesizey))) == NULL)
 				_throwunix("allocating image buffers");
 		}

 		// Compression test
-		if(quiet) printf("%s\t%s\t%s\t%d\t",  _pfname[pf], bu?"BU":"TD",
+		if(quiet==1) printf("%s\t%s\t%s\t%d\t",  _pfname[pf], bu?"BU":"TD",
 			_subnamel[jpegsub], qual);
 		for(i=0; i<h; i++) memcpy(&rgbbuf[pitch*i], &srcbuf[w*ps*i], w*ps);
 		if((hnd=tjInitCompress())==NULL)
 			_throwtj("executing tjInitCompress()");
-		if(tjCompress(hnd, rgbbuf, tilesizex, pitch, tilesizey, ps,
+		if(yuv==YUVENCODE)
+		{
+			if(tjEncodeYUV(hnd, rgbbuf, tilesizex, pitch, tilesizey, ps,
+				jpegbuf[0], jpegsub, flags)==-1)
+				_throwtj("executing tjEncodeYUV()");
+			comptilesize[0]=TJBUFSIZEYUV(tilesizex, tilesizey, jpegsub);
+		}
+		else if(tjCompress(hnd, rgbbuf, tilesizex, pitch, tilesizey, ps,
 			jpegbuf[0], &comptilesize[0], jpegsub, qual, flags)==-1)
 			_throwtj("executing tjCompress()");
 		ITER=0;
-		timer.start();
+		start=rrtime();
 		do
 		{
-			jpgbufsize=0;  int tilen=0;
+			jpgbufsize=0;  tilen=0;
 			for(i=0; i<h; i+=tilesizey)
 			{
 				for(j=0; j<w; j+=tilesizex)
 				{
 					int tempw=min(tilesizex, w-j), temph=min(tilesizey, h-i);
-					if(tjCompress(hnd, &rgbbuf[pitch*i+j*ps], tempw, pitch,
+					if(yuv==YUVENCODE)
+					{
+						if(tjEncodeYUV(hnd, &rgbbuf[pitch*i+j*ps], tempw, pitch,
+							temph, ps, jpegbuf[tilen], jpegsub, flags)==-1)
+							_throwtj("executing tjEncodeYUV()");
+						comptilesize[tilen]=TJBUFSIZEYUV(tempw, temph, jpegsub);
+					}
+					else if(tjCompress(hnd, &rgbbuf[pitch*i+j*ps], tempw, pitch,
 						temph, ps, jpegbuf[tilen], &comptilesize[tilen], jpegsub, qual,
 						flags)==-1)
 						_throwtj("executing tjCompress()");
@@ -147,17 +157,20 @@ void dotest(unsigned char *srcbuf, int w, int h, BMPPIXELFORMAT pf, int bu,
 				}
 			}
 			ITER++;
-		} while((elapsed=timer.elapsed())<5.);
+		} while((elapsed=rrtime()-start)<5.);
 		if(tjDestroy(hnd)==-1) _throwtj("executing tjDestroy()");
 		hnd=NULL;
-		if(quiet)
+		if(quiet==1)
 		{
 			if(tilesizex==w && tilesizey==h) printf("Full     \t");
 			else printf("%-4d %-4d\t", tilesizex, tilesizey);
+		}
+		if(quiet)
+		{
 			printsigfig((double)(w*h)/1000000.*(double)ITER/elapsed, 4);
-			printf("\t");
+			printf("%c", quiet==2? '\n':'\t');
 			printsigfig((double)(w*h*ps)/(double)jpgbufsize, 4);
-			printf("\t");
+			printf("%c", quiet==2? '\n':'\t');
 		}
 		else
 		{
@@ -185,18 +198,25 @@ void dotest(unsigned char *srcbuf, int w, int h, BMPPIXELFORMAT pf, int bu,
 			fclose(outfile);  outfile=NULL;
 			if(!quiet) printf("Reference image written to %s\n", tempstr);
 		}
-		if(yuv==YUVENCODE) goto bailout;
+		if(yuv==YUVENCODE)
+		{
+			if(quiet==1) printf("\n");  goto bailout;
+		}

 		// Decompression test
-		if(yuv==YUVDECODE) flags |= TJ_YUV;
 		memset(rgbbuf, 127, max(yuvsize, pitch*h));  // Grey image means decompressor did nothing
 		if((hnd=tjInitDecompress())==NULL)
 			_throwtj("executing tjInitDecompress()");
-		if(tjDecompress(hnd, jpegbuf[0], jpgbufsize, rgbbuf, tilesizex, pitch,
+		if(yuv==YUVDECODE)
+		{
+			if(tjDecompressToYUV(hnd, jpegbuf[0], jpgbufsize, rgbbuf, flags)==-1)
+				_throwtj("executing tjDecompressToYUV()");
+		}
+		else if(tjDecompress(hnd, jpegbuf[0], jpgbufsize, rgbbuf, tilesizex, pitch,
 			tilesizey, ps, flags)==-1)
 			_throwtj("executing tjDecompress()");
 		ITER=0;
-		timer.start();
+		start=rrtime();
 		do
 		{
 			int tilen=0;
@@ -205,14 +225,20 @@ void dotest(unsigned char *srcbuf, int w, int h, BMPPIXELFORMAT pf, int bu,
 				for(j=0; j<w; j+=tilesizex)
 				{
 					int tempw=min(tilesizex, w-j), temph=min(tilesizey, h-i);
-					if(tjDecompress(hnd, jpegbuf[tilen], comptilesize[tilen],
+					if(yuv==YUVDECODE)
+					{
+						if(tjDecompressToYUV(hnd, jpegbuf[tilen], comptilesize[tilen],
+							&rgbbuf[pitch*i+ps*j], flags)==-1)
+							_throwtj("executing tjDecompressToYUV()");
+					}
+					else if(tjDecompress(hnd, jpegbuf[tilen], comptilesize[tilen],
 						&rgbbuf[pitch*i+ps*j], tempw, pitch, temph, ps, flags)==-1)
 						_throwtj("executing tjDecompress()");
 					tilen++;
 				}
 			}
 			ITER++;
-		}	while((elapsed=timer.elapsed())<5.);
+		}	while((elapsed=rrtime()-start)<5.);
 		if(tjDestroy(hnd)==-1) _throwtj("executing tjDestroy()");
 		hnd=NULL;
 		if(quiet)
@@ -301,13 +327,13 @@ void dotest(unsigned char *srcbuf, int w, int h, BMPPIXELFORMAT pf, int bu,
 }


-void dodecomptest(char *filename, BMPPIXELFORMAT pf, int bu, int useppm,
+void dodecomptest(char *filename, int pf, int bu, int useppm,
 	int quiet)
 {
 	char tempstr[1024];
-	FILE *file=NULL;  tjhandle hnd;
+	FILE *file=NULL;  tjhandle hnd=NULL;
 	unsigned char *jpegbuf=NULL, *rgbbuf=NULL;
-	rrtimer timer; double elapsed;
+	double start, elapsed;
 	int w, h, ITER;
 	unsigned long jpgbufsize=0;
 	int flags=(forcemmx?TJ_FORCEMMX:0)|(forcesse?TJ_FORCESSE:0)
@@ -315,11 +341,10 @@ void dodecomptest(char *filename, BMPPIXELFORMAT pf, int bu, int useppm,
 		|(fastupsample?TJ_FASTUPSAMPLE:0);
 	int ps=_ps[pf], pitch, jpegsub=-1;
 	char *temp=NULL;
-	int hsf, vsf, pw, ph, cw, ch, ypitch, uvpitch, yuvsize;
+	int yuvsize, bufsize;

 	flags |= _flags[pf];
 	if(bu) flags |= TJ_BOTTOMUP;
-	if(yuv==YUVDECODE) flags |= TJ_YUV;

 	if((file=fopen(filename, "rb"))==NULL)
 		_throwunix("opening file");
@@ -340,15 +365,11 @@ void dodecomptest(char *filename, BMPPIXELFORMAT pf, int bu, int useppm,
 	if(tjDecompressHeader2(hnd, jpegbuf, jpgbufsize, &w, &h, &jpegsub)==-1)
 		_throwtj("executing tjDecompressHeader2()");

-	hsf=_hsf[jpegsub], vsf=_vsf[jpegsub];
-	pw=PAD(w, hsf), ph=PAD(h, vsf);
-	cw=pw/hsf, ch=ph/vsf;
-	ypitch=PAD(pw, 4), uvpitch=PAD(cw, 4);
-	yuvsize=ypitch*ph + (jpegsub==TJ_GRAYSCALE? 0:uvpitch*ch*2);
+	yuvsize=TJBUFSIZEYUV(w, h, jpegsub);

 	pitch=w*ps;

-	if(quiet)
+	if(quiet==1)
 	{
 		printf("All performance values in Mpixels/sec\n\n");
 		printf("Bitmap\tBitmap\tImage Size\tDecomp\n"),
@@ -356,31 +377,43 @@ void dodecomptest(char *filename, BMPPIXELFORMAT pf, int bu, int useppm,
 		printf("%s\t%s\t%-4d %-4d\t", _pfname[pf], bu?"BU":"TD", w, h);
 	}

-	if((rgbbuf=(unsigned char *)malloc(max(yuvsize, pitch*h)))==NULL)
+	bufsize=(yuv==YUVDECODE? yuvsize:pitch*h);
+	if((rgbbuf=(unsigned char *)malloc(bufsize))==NULL)
 		_throwunix("allocating image buffer");

 	if(!quiet)
 	{
 		if(yuv==YUVDECODE)
-			printf("\n>>>>>  JPEG --> YUV %s  <<<<<\n", _subnamel[jpegsub]);
+			printf(">>>>>  JPEG --> YUV %s  <<<<<\n", _subnamel[jpegsub]);
 		else
-			printf("\n>>>>>  JPEG --> %s (%s)  <<<<<\n", _pfname[pf],
+			printf(">>>>>  JPEG --> %s (%s)  <<<<<\n", _pfname[pf],
 				bu?"Bottom-up":"Top-down");
 		printf("\nImage size: %d x %d\n", w, h);
 	}

-	memset(rgbbuf, 127, max(yuvsize, pitch*h));  // Grey image means decompressor did nothing
-	if(tjDecompress(hnd, jpegbuf, jpgbufsize, rgbbuf, w, pitch, h, ps, flags)==-1)
+	memset(rgbbuf, 127, bufsize);  // Grey image means decompressor did nothing
+	if(yuv==YUVDECODE)
+	{
+		if(tjDecompressToYUV(hnd, jpegbuf, jpgbufsize, rgbbuf, flags)==-1)
+			_throwtj("executing tjDecompressToYUV()");
+	}
+	else if(tjDecompress(hnd, jpegbuf, jpgbufsize, rgbbuf, w, pitch, h, ps,
+		flags)==-1)
 		_throwtj("executing tjDecompress()");
 	ITER=0;
-	timer.start();
+	start=rrtime();
 	do
 	{
-		if(tjDecompress(hnd, jpegbuf, jpgbufsize, rgbbuf, w, pitch, h, ps, flags)
-			==-1)
+		if(yuv==YUVDECODE)
+		{
+			if(tjDecompressToYUV(hnd, jpegbuf, jpgbufsize, rgbbuf, flags)==-1)
+				_throwtj("executing tjDecompressToYUV()");
+		}
+		else if(tjDecompress(hnd, jpegbuf, jpgbufsize, rgbbuf, w, pitch, h, ps,
+			flags)==-1)
 			_throwtj("executing tjDecompress()");
 		ITER++;
-	}	while((elapsed=timer.elapsed())<5.);
+	}	while((elapsed=rrtime()-start)<5.);
 	if(tjDestroy(hnd)==-1) _throwtj("executing tjDestroy()");
 	hnd=NULL;
 	if(quiet)
@@ -427,8 +460,8 @@ void usage(char *progname)
 	printf("       Test performance of the codec when the image is encoded\n");
 	printf("       as separate tiles of varying sizes.\n\n");
 	printf("       [-forcemmx] [-forcesse] [-forcesse2] [-forcesse3]\n");
-	printf("       Force MMX, SSE, or SSE2 code paths in Intel codec\n\n");
-	printf("       [-rgb | -bgr | -rgba | -bgra | -abgr | -argb]\n");
+	printf("       Force MMX, SSE, SSE2, or SSE3 code paths in the underlying codec\n\n");
+	printf("       [-rgb | -bgr | -rgbx | -bgrx | -xbgr | -xrgb]\n");
 	printf("       Test the specified color conversion path in the codec (default: BGR)\n\n");
 	printf("       [-fastupsample]\n");
 	printf("       Use fast, inaccurate upsampling code to perform 4:2:2 and 4:2:0\n");
@@ -448,12 +481,10 @@ void usage(char *progname)
 int main(int argc, char *argv[])
 {
 	unsigned char *bmpbuf=NULL;  int w, h, i, useppm=0;
-	int qual, dotile=0, quiet=0, hiqual=-1;  char *temp;
-	BMPPIXELFORMAT pf=BMP_BGR;
+	int qual=-1, dotile=0, quiet=0, hiqual=-1;  char *temp;
+	int pf=BMP_BGR;
 	int bu=0, minarg=2;

-	printf("\n");
-
 	if(argc<minarg) usage(argv[0]);

 	temp=strrchr(argv[1], '.');
@@ -463,18 +494,20 @@ int main(int argc, char *argv[])
 		if(!stricmp(temp, ".jpg") || !stricmp(temp, ".jpeg")) decomponly=1;
 	}

+	printf("\n");
+
 	if(argc>minarg)
 	{
 		for(i=minarg; i<argc; i++)
 		{
 			if(!stricmp(argv[i], "-yuvencode"))
 			{
-				printf("Testing YUV planar encoding\n");
+				printf("Testing YUV planar encoding\n\n");
 				yuv=YUVENCODE;  hiqual=qual=100;
 			}
 			if(!stricmp(argv[i], "-yuvdecode"))
 			{
-				printf("Testing YUV planar decoding\n");
+				printf("Testing YUV planar decoding\n\n");
 				yuv=YUVDECODE;
 			}
 		}
@@ -502,37 +535,38 @@ int main(int argc, char *argv[])
 			if(!stricmp(argv[i], "-tile")) dotile=1;
 			if(!stricmp(argv[i], "-forcesse3"))
 			{
-				printf("Using SSE3 code\n");
+				printf("Using SSE3 code\n\n");
 				forcesse3=1;
 			}
 			if(!stricmp(argv[i], "-forcesse2"))
 			{
-				printf("Using SSE2 code\n");
+				printf("Using SSE2 code\n\n");
 				forcesse2=1;
 			}
 			if(!stricmp(argv[i], "-forcesse"))
 			{
-				printf("Using SSE code\n");
+				printf("Using SSE code\n\n");
 				forcesse=1;
 			}
 			if(!stricmp(argv[i], "-forcemmx"))
 			{
-				printf("Using MMX code\n");
+				printf("Using MMX code\n\n");
 				forcemmx=1;
 			}
 			if(!stricmp(argv[i], "-fastupsample"))
 			{
-				printf("Using fast upsampling code\n");
+				printf("Using fast upsampling code\n\n");
 				fastupsample=1;
 			}
 			if(!stricmp(argv[i], "-rgb")) pf=BMP_RGB;
-			if(!stricmp(argv[i], "-rgba")) pf=BMP_RGBA;
+			if(!stricmp(argv[i], "-rgbx")) pf=BMP_RGBX;
 			if(!stricmp(argv[i], "-bgr")) pf=BMP_BGR;
-			if(!stricmp(argv[i], "-bgra")) pf=BMP_BGRA;
-			if(!stricmp(argv[i], "-abgr")) pf=BMP_ABGR;
-			if(!stricmp(argv[i], "-argb")) pf=BMP_ARGB;
+			if(!stricmp(argv[i], "-bgrx")) pf=BMP_BGRX;
+			if(!stricmp(argv[i], "-xbgr")) pf=BMP_XBGR;
+			if(!stricmp(argv[i], "-xrgb")) pf=BMP_XRGB;
 			if(!stricmp(argv[i], "-bottomup")) bu=1;
 			if(!stricmp(argv[i], "-quiet")) quiet=1;
+			if(!stricmp(argv[i], "-qq")) quiet=2;
 		}
 	}

@@ -544,7 +578,7 @@ int main(int argc, char *argv[])
 		if(temp!=NULL) *temp='\0';
 	}

-	if(quiet && !decomponly)
+	if(quiet==1 && !decomponly)
 	{
 		printf("All performance values in Mpixels/sec\n\n");
 		printf("Bitmap\tBitmap\tJPEG\tJPEG\tTile Size\tCompr\tCompr\tDecomp\n");
@@ -554,19 +588,21 @@ int main(int argc, char *argv[])
 	if(decomponly)
 	{
 		dodecomptest(argv[1], pf, bu, 1, quiet);
+		printf("\n");
 		goto bailout;
 	}
 	for(i=hiqual; i>=qual; i--)
 		dotest(bmpbuf, w, h, pf, bu, TJ_GRAYSCALE, i, argv[1], dotile, useppm, quiet);
-	if(quiet) printf("\n");
+	printf("\n");
 	for(i=hiqual; i>=qual; i--)
 		dotest(bmpbuf, w, h, pf, bu, TJ_420, i, argv[1], dotile, useppm, quiet);
-	if(quiet) printf("\n");
+	printf("\n");
 	for(i=hiqual; i>=qual; i--)
 		dotest(bmpbuf, w, h, pf, bu, TJ_422, i, argv[1], dotile, useppm, quiet);
-	if(quiet) printf("\n");
+	printf("\n");
 	for(i=hiqual; i>=qual; i--)
 		dotest(bmpbuf, w, h, pf, bu, TJ_444, i, argv[1], dotile, useppm, quiet);
+	printf("\n");

 	bailout:
 	if(bmpbuf) free(bmpbuf);
--- a/jsimd.h
+++ b/jsimd.h
@@ -5,6 +5,7 @@
 * 
 * Based on the x86 SIMD extension for IJG JPEG library,
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
+ * For conditions of distribution and use, see copyright notice in jsimdext.inc
 *
 */

--- a/jsimd_none.c
+++ b/jsimd_none.c
@@ -6,6 +6,7 @@
 * 
 * Based on the x86 SIMD extension for IJG JPEG library,
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
+ * For conditions of distribution and use, see copyright notice in jsimdext.inc
 *
 * This file contains stubs for when there is no SIMD support available.
 */
--- a/jsimddct.h
+++ b/jsimddct.h
@@ -5,6 +5,7 @@
 * 
 * Based on the x86 SIMD extension for IJG JPEG library,
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
+ * For conditions of distribution and use, see copyright notice in jsimdext.inc
 *
 */

--- a/jversion.h
+++ b/jversion.h
@@ -34,4 +34,4 @@
 			"Copyright (C) 2004 Landmark Graphics Corporation\n" \
 			"Copyright (C) 2005-2007 Sun Microsystems, Inc.\n" \
 			"Copyright (C) 2009 Pierre Ossman for Cendio AB\n" \
-			"Copyright (C) 2009-2010 D. R. Commander"
+			"Copyright (C) 2009-2011 D. R. Commander"
--- a/release/ReadMe.rtf
+++ b/release/ReadMe.rtf
@@ -1,11 +1,13 @@
-{\rtf1\ansi\ansicpg1252\cocoartf1038\cocoasubrtf250
+{\rtf1\ansi\ansicpg1252\cocoartf1038\cocoasubrtf350
 {\fonttbl\f0\fswiss\fcharset0 Helvetica;}
 {\colortbl;\red255\green255\blue255;}
 \margl1440\margr1440\vieww15200\viewh9600\viewkind0
 \deftab720
 \pard\pardeftab720\ql\qnatural

-\f0\fs24 \cf0 libjpeg-turbo is a high-speed version of libjpeg for x86 and x86-64 processors which uses SIMD instructions (MMX, SSE2, etc.) to accelerate baseline JPEG compression and decompression.  libjpeg-turbo is generally 2-4x as fast as the unmodified version of libjpeg, all else being equal.  libjpeg-turbo also includes a wrapper library for the TurboJPEG API used by VirtualGL and TurboVNC.\
+\f0\fs24 \cf0 libjpeg-turbo is a derivative of libjpeg which uses SIMD instructions (MMX, SSE2, etc.) to accelerate baseline JPEG compression and decompression on x86 and x86-64 systems.  On such systems, libjpeg-turbo is generally 2-4x as fast as the unmodified version of libjpeg, all else being equal.  libjpeg-turbo also includes a wrapper library which implements the TurboJPEG API used by VirtualGL and TurboVNC.\
 \
-libjpeg-turbo was originally based on libjpeg/SIMD by Miyasaka Masaru, but the TigerVNC and VirtualGL projects made numerous enhancements to the codec, including improved support for Mac OS X, 64-bit support, support for 32-bit and big endian pixel formats, accelerated Huffman encoding/decoding, and various bug fixes.  The goal was to produce a fully open source codec that could replace the partially closed source TurboJPEG/IPP codec used by VirtualGL and TurboVNC.  libjpeg-turbo generally performs in the range of 80-120% of TurboJPEG/IPP.  It is faster in some areas but slower in others.\
+libjpeg-turbo was originally based on libjpeg/SIMD by Miyasaka Masaru, but the TigerVNC and VirtualGL projects made numerous enhancements to the codec in 2009, including improved support for Mac OS X, 64-bit support, support for 32-bit and big endian pixel formats (RGBX, XBGR, etc.), accelerated Huffman encoding/decoding, and various bug fixes.  The goal was to produce a fully open source codec that could replace the partially closed source TurboJPEG/IPP codec used by VirtualGL and TurboVNC.  libjpeg-turbo generally performs in the range of 80-120% of TurboJPEG/IPP.  It is faster in some areas but slower in others.\
+\
+In early 2010, libjpeg-turbo spun off into its own independent project, with the goal of making high-speed JPEG compression/decompression technology available to a broader range of users and developers.\
 }
--- a/release/deb-control.tmpl
+++ b/release/deb-control.tmpl
@@ -5,19 +5,24 @@ Priority: optional
 Architecture: {__ARCH}
 Essential: no
 Maintainer: The libjpeg-turbo Project [http://libjpeg-turbo.VirtualGL.org]
-Description: A toolkit for displaying OpenGL applications to thin clients
- libjpeg-turbo is a high-speed version of libjpeg for x86 and x86-64 processors
- which uses SIMD instructions (MMX, SSE2, etc.) to accelerate baseline JPEG
- compression and decompression.  libjpeg-turbo is generally 2-4x as fast
+Description: A SIMD-accelerated JPEG codec which provides both the libjpeg and TurboJPEG APIs
+ libjpeg-turbo is a derivative of libjpeg which uses SIMD instructions (MMX,
+ SSE2, etc.) to accelerate baseline JPEG compression and decompression on x86
+ and x86-64 systems.  On such systems, libjpeg-turbo is generally 2-4x as fast
 as the unmodified version of libjpeg, all else being equal.  libjpeg-turbo
- also includes a wrapper library for the TurboJPEG API used by VirtualGL and
- TurboVNC.
+ also includes a wrapper library which implements the TurboJPEG API used by
+ VirtualGL and TurboVNC.
 .
 libjpeg-turbo was originally based on libjpeg/SIMD by Miyasaka Masaru, but
- the TigerVNC and VirtualGL projects made numerous enhancements to the codec,
- including improved support for Mac OS X, 64-bit support, support for 32-bit
- and big endian pixel formats, accelerated Huffman encoding/decoding, and
- various bug fixes.  The goal was to produce a fully open source codec that
- could replace the partially closed source TurboJPEG/IPP codec used by
- VirtualGL and TurboVNC.  libjpeg-turbo generally performs in the range of
- 80-120% of TurboJPEG/IPP.  It is faster in some areas but slower in others.
+ the TigerVNC and VirtualGL projects made numerous enhancements to the codec
+ in 2009, including improved support for Mac OS X, 64-bit support, support
+ for 32-bit and big endian pixel formats (RGBX, XBGR, etc.), accelerated
+ Huffman encoding/decoding, and various bug fixes.  The goal was to produce a
+ fully open source codec that could replace the partially closed source
+ TurboJPEG/IPP codec used by VirtualGL and TurboVNC.  libjpeg-turbo generally
+ performs in the range of 80-120% of TurboJPEG/IPP.  It is faster in some
+ areas but slower in others.
+ .
+ In early 2010, libjpeg-turbo spun off into its own independent project, with
+ the goal of making high-speed JPEG compression/decompression technology
+ available to a broader range of users and developers.
--- a/release/libjpeg-turbo.nsi.in
+++ b/release/libjpeg-turbo.nsi.in
@@ -45,6 +45,12 @@ Section "@CMAKE_PROJECT_NAME@ SDK for @INST_PLATFORM@ (required)"
 !else
 	File "@CMAKE_BINARY_DIR@\sharedlib\${BUILDDIR}jpeg@DLL_VERSION@.dll"
 !endif
+	File "@CMAKE_BINARY_DIR@\sharedlib\cjpeg.exe"
+	File "@CMAKE_BINARY_DIR@\sharedlib\djpeg.exe"
+	File "@CMAKE_BINARY_DIR@\sharedlib\jpegtran.exe"
+	File "@CMAKE_BINARY_DIR@\jpgtest.exe"
+	File "@CMAKE_BINARY_DIR@\rdjpgcom.exe"
+	File "@CMAKE_BINARY_DIR@\wrjpgcom.exe"
 	SetOutPath $INSTDIR\lib
 !ifdef GCC
 	File "@CMAKE_BINARY_DIR@\libturbojpeg.dll.a"
@@ -63,12 +69,13 @@ Section "@CMAKE_PROJECT_NAME@ SDK for @INST_PLATFORM@ (required)"
 	File "@CMAKE_SOURCE_DIR@\jmorecfg.h"
 	File "@CMAKE_SOURCE_DIR@\jpeglib.h"
 	File "@CMAKE_SOURCE_DIR@\turbojpeg.h"
-	SetOutPath $INSTDIR
+	SetOutPath $INSTDIR\doc
 	File "@CMAKE_SOURCE_DIR@\README"
 	File "@CMAKE_SOURCE_DIR@\README-turbo.txt"
 	File "@CMAKE_SOURCE_DIR@\libjpeg.txt"
 	File "@CMAKE_SOURCE_DIR@\LGPL.txt"
 	File "@CMAKE_SOURCE_DIR@\LICENSE.txt"
+	File "@CMAKE_SOURCE_DIR@\usage.txt"

 	WriteRegStr HKLM "SOFTWARE\@INST_NAME@ @VERSION@" "Install_Dir" "$INSTDIR"

@@ -106,20 +113,28 @@ Section "Uninstall"
 	Delete $INSTDIR\lib\turbojpeg.lib
 	Delete $INSTDIR\lib\turbojpeg-static.lib
 !endif
+	Delete $INSTDIR\bin\cjpeg.exe
+	Delete $INSTDIR\bin\djpeg.exe
+	Delete $INSTDIR\bin\jpegtran.exe
+	Delete $INSTDIR\bin\jpgtest.exe
+	Delete $INSTDIR\bin\rdjpgcom.exe
+	Delete $INSTDIR\bin\wrjpgcom.exe
 	Delete $INSTDIR\include\jconfig.h"
 	Delete $INSTDIR\include\jerror.h"
 	Delete $INSTDIR\include\jmorecfg.h"
 	Delete $INSTDIR\include\jpeglib.h"
 	Delete $INSTDIR\include\turbojpeg.h"
 	Delete $INSTDIR\uninstall_@VERSION@.exe
-	Delete $INSTDIR\README
-	Delete $INSTDIR\README-turbo.txt
-	Delete $INSTDIR\libjpeg.txt
-	Delete $INSTDIR\LGPL.txt
-	Delete $INSTDIR\LICENSE.txt
+	Delete $INSTDIR\doc\README
+	Delete $INSTDIR\doc\README-turbo.txt
+	Delete $INSTDIR\doc\libjpeg.txt
+	Delete $INSTDIR\doc\LGPL.txt
+	Delete $INSTDIR\doc\LICENSE.txt
+	Delete $INSTDIR\doc\usage.txt

 	RMDir "$INSTDIR\include"
 	RMDir "$INSTDIR\lib"
+	RMDir "$INSTDIR\doc"
 	RMDir "$INSTDIR\bin"
 	RMDir "$INSTDIR"

--- a/release/libjpeg-turbo.spec.in
+++ b/release/libjpeg-turbo.spec.in
@@ -8,7 +8,7 @@ Summary: A SIMD-accelerated JPEG codec which provides both the libjpeg and Turbo
 Name: @PACKAGE_NAME@
 Version: @VERSION@
 Vendor: The libjpeg-turbo Project
-URL: http://libjpeg-turbo.virtualgl.org
+URL: http://libjpeg-turbo.VirtualGL.org
 Group: System Environment/Libraries
 #-->Source0: http://prdownloads.sourceforge.net/libjpeg-turbo/libjpeg-turbo-%{version}.tar.gz
 Release: @BUILD@
@@ -19,21 +19,25 @@ Provides: %{name} = %{version}-%{release}, turbojpeg = 2.00
 Obsoletes: turbojpeg

 %description
-libjpeg-turbo is a high-speed version of libjpeg for x86 and x86-64 processors
-which uses SIMD instructions (MMX, SSE2, etc.) to accelerate baseline JPEG
-compression and decompression.  libjpeg-turbo is generally 2-4x as fast
+libjpeg-turbo is a derivative of libjpeg which uses SIMD instructions (MMX,
+SSE2, etc.) to accelerate baseline JPEG compression and decompression on x86
+and x86-64 systems.  On such systems, libjpeg-turbo is generally 2-4x as fast
 as the unmodified version of libjpeg, all else being equal.  libjpeg-turbo also
-includes a wrapper library for the TurboJPEG API used by VirtualGL and
-TurboVNC.
+includes a wrapper library which implements the TurboJPEG API used by VirtualGL
+and TurboVNC.

 libjpeg-turbo was originally based on libjpeg/SIMD by Miyasaka Masaru, but
-the TigerVNC and VirtualGL projects made numerous enhancements to the codec,
-including improved support for Mac OS X, 64-bit support, support for 32-bit
-and big endian pixel formats, accelerated Huffman encoding/decoding, and
-various bug fixes.  The goal was to produce a fully open source codec that
-could replace the partially closed source TurboJPEG/IPP codec used by VirtualGL
-and TurboVNC.  libjpeg-turbo generally performs in the range of 80-120% of
-TurboJPEG/IPP.  It is faster in some areas but slower in others.
+the TigerVNC and VirtualGL projects made numerous enhancements to the codec in
+2009, including improved support for Mac OS X, 64-bit support, support for
+32-bit and big endian pixel formats (RGBX, XBGR, etc.), accelerated Huffman
+encoding/decoding, and various bug fixes.  The goal was to produce a fully open
+source codec that could replace the partially closed source TurboJPEG/IPP codec
+used by VirtualGL and TurboVNC.  libjpeg-turbo generally performs in the range
+of 80-120% of TurboJPEG/IPP.  It is faster in some areas but slower in others.
+
+In early 2010, libjpeg-turbo spun off into its own independent project, with
+the goal of making high-speed JPEG compression/decompression technology
+available to a broader range of users and developers.

 #-->%prep
 #-->%setup -q
@@ -71,6 +75,7 @@ rm -rf $RPM_BUILD_ROOT
 /opt/%{name}/bin/cjpeg
 /opt/%{name}/bin/djpeg
 /opt/%{name}/bin/jpegtran
+/opt/%{name}/bin/jpgtest
 /opt/%{name}/bin/rdjpgcom
 /opt/%{name}/bin/wrjpgcom
 %dir /opt/%{name}/%{__lib}
--- a/release/makemacpkg.in
+++ b/release/makemacpkg.in
@@ -102,6 +102,10 @@ if [ $UNIVERSAL = 1 ]; then
 		-arch i386 $TMPDIR/dist.x86/opt/$PACKAGE_NAME/bin/jpegtran \
 		-arch x86_64 $PKGROOT/opt/$PACKAGE_NAME/bin/jpegtran \
 		-output $PKGROOT/opt/$PACKAGE_NAME/bin/jpegtran
+	lipo -create \
+		-arch i386 $TMPDIR/dist.x86/opt/$PACKAGE_NAME/bin/jpgtest \
+		-arch x86_64 $PKGROOT/opt/$PACKAGE_NAME/bin/jpgtest \
+		-output $PKGROOT/opt/$PACKAGE_NAME/bin/jpgtest
 	lipo -create \
 		-arch i386 $TMPDIR/dist.x86/opt/$PACKAGE_NAME/bin/rdjpgcom \
 		-arch x86_64 $PKGROOT/opt/$PACKAGE_NAME/bin/rdjpgcom \
--- a/release/makesunpkg.in
+++ b/release/makesunpkg.in
@@ -81,6 +81,7 @@ d none $PACKAGE_NAME/bin 0755 root bin
 f none $PACKAGE_NAME/bin/cjpeg 0755 root bin
 f none $PACKAGE_NAME/bin/djpeg 0755 root bin
 f none $PACKAGE_NAME/bin/jpegtran 0755 root bin
+f none $PACKAGE_NAME/bin/jpgtest 0755 root bin
 f none $PACKAGE_NAME/bin/rdjpgcom 0755 root bin
 f none $PACKAGE_NAME/bin/wrjpgcom 0755 root bin
 d none $PACKAGE_NAME/lib 0755 root bin
--- a/rrtimer.h
+++ b/rrtimer.h
@@ -15,69 +15,6 @@
 #ifndef __RRTIMER_H__
 #define __RRTIMER_H__

-#ifdef __cplusplus
-
-#ifdef _WIN32
-#include <windows.h>
-#else
-#include <sys/time.h>
-#endif
-
-class rrtimer
-{
-	public:
-
-		rrtimer(void) : t1(0.0)
-		{
-			#ifdef _WIN32
-			highres=false;  tick=0.001;
-			LARGE_INTEGER Frequency;
-			if(QueryPerformanceFrequency(&Frequency)!=0)
-			{
-				tick=(double)1.0/(double)(Frequency.QuadPart);
-				highres=true;
-			}
-			#endif
-		}
-
-		void start(void)
-		{
-			t1=time();
-		}
-
-		double time(void)
-		{
-			#ifdef _WIN32
-			if(highres)
-			{
-				LARGE_INTEGER Time;
-				QueryPerformanceCounter(&Time);
-				return((double)(Time.QuadPart)*tick);
-			}
-			else
-				return((double)GetTickCount()*tick);
-			#else
-			struct timeval __tv;
-			gettimeofday(&__tv, (struct timezone *)NULL);
-			return((double)(__tv.tv_sec)+(double)(__tv.tv_usec)*0.000001);
-			#endif
-		}
-
-		double elapsed(void)
-		{
-			return time()-t1;
-		}
-
-	private:
-
-		#ifdef _WIN32
-		bool highres;  double tick;
-		#endif
-		double t1;
-};
-
-#endif  // __cplusplus
-
 #ifdef _WIN32

 #include <windows.h>
@@ -111,4 +48,3 @@ static __inline double rrtime(void)
 #endif

 #endif
-
--- a/sharedlib/CMakeLists.txt
+++ b/sharedlib/CMakeLists.txt
@@ -8,9 +8,7 @@
 if(MSVC)
  # Build all configurations against shared C library
  foreach(var CMAKE_C_FLAGS CMAKE_C_FLAGS_DEBUG CMAKE_C_FLAGS_RELEASE
-    CMAKE_C_FLAGS_MINSIZEREL CMAKE_C_FLAGS_RELWITHDEBINFO
-    CMAKE_CXX_FLAGS CMAKE_CXX_FLAGS_DEBUG CMAKE_CXX_FLAGS_RELEASE
-    CMAKE_CXX_FLAGS_MINSIZEREL CMAKE_CXX_FLAGS_RELWITHDEBINFO)
+    CMAKE_C_FLAGS_MINSIZEREL CMAKE_C_FLAGS_RELWITHDEBINFO)
    if(${var} MATCHES "/MT")
      string(REGEX REPLACE "/MT" "/MD" ${var} "${${var}}")
    endif()
--- a/simd/jsimd.h
+++ b/simd/jsimd.h
@@ -5,6 +5,7 @@
 * 
 * Based on the x86 SIMD extension for IJG JPEG library,
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
+ * For conditions of distribution and use, see copyright notice in jsimdext.inc
 *
 */

--- a/simd/jsimd_i386.c
+++ b/simd/jsimd_i386.c
@@ -6,6 +6,7 @@
 * 
 * Based on the x86 SIMD extension for IJG JPEG library,
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
+ * For conditions of distribution and use, see copyright notice in jsimdext.inc
 *
 * This file contains the interface between the "normal" portions
 * of the library and the SIMD implementations when running on a
--- a/simd/jsimd_x86_64.c
+++ b/simd/jsimd_x86_64.c
@@ -6,6 +6,7 @@
 * 
 * Based on the x86 SIMD extension for IJG JPEG library,
 * Copyright (C) 1999-2006, MIYASAKA Masaru.
+ * For conditions of distribution and use, see copyright notice in jsimdext.inc
 *
 * This file contains the interface between the "normal" portions
 * of the library and the SIMD implementations when running on a
--- a/testimgfst100.jpg
+++ b/testimgfst100.jpg
--- a/11
+++ b/11
@@ -1,3 +1,4 @@
+TURBOJPEG_1.0
 {
 	global:
 		tjInitCompress;
@@ -5,10 +6,18 @@
 		TJBUFSIZE;
 		tjInitDecompress;
 		tjDecompressHeader;
-		tjDecompressHeader2;
 		tjDecompress;
 		tjDestroy;
 		tjGetErrorStr;
 	local:
 		*;
 };
+
+TURBOJPEG_1.1
+{
+	global:
+		TJBUFSIZEYUV;
+		tjDecompressHeader2;
+		tjDecompressToYUV;
+		tjEncodeYUV;
+} TURBOJPEG_1.0;
--- a/turbojpeg.h
+++ b/turbojpeg.h
@@ -1,6 +1,6 @@
 /* Copyright (C)2004 Landmark Graphics Corporation
 * Copyright (C)2005, 2006 Sun Microsystems, Inc.
- * Copyright (C)2009 D. R. Commander
+ * Copyright (C)2009-2011 D. R. Commander
 *
 * This library is free software and may be redistributed and/or modified under
 * the terms of the wxWindows Library License, Version 3.1 or (at your option)
@@ -52,28 +52,9 @@ enum {TJ_444=0, TJ_422, TJ_420, TJ_GRAYSCALE};
     (64-bit IPP version only) */
 #define TJ_FASTUPSAMPLE  256
  /* Use fast, inaccurate 4:2:2 and 4:2:0 YUV upsampling routines
-     (libjpeg version only) */
+     (libjpeg and libjpeg-turbo versions only) */
 #define TJ_YUV           512
-  /* If passed to tjCompress(), this causes TurboJPEG/OSS to use the
-     accelerated color conversion routines in libjpeg-turbo to produce a planar
-     YUV image that is suitable for X Video.  Specifically, if the chrominance
-     components are subsampled along the horizontal dimension, then the width
-     of the luminance plane is padded to 2 in the output image (same goes for
-     the height of the luminance plane, if the chrominance components are
-     subsampled along the vertical dimension.)  Also, each line of each plane
-     in the output image is padded to 4 bytes.  Although this will work with
-     any subsampling option, it is really only useful in combination with
-     TJ_420, which produces an image compatible with the I420 (AKA "YUV420P")
-     format.
-
-     If passed to tjDecompress(), this tells TurboJPEG/OSS to perform JPEG
-     decompression but to leave out the color conversion step, so a planar YUV
-     image is generated instead of an RGB image.  The padding of the planes in
-     this image is the same as in the above case.  Note that, if the width or
-     height of the output image is not a multiple of 8 (or a multiple of 16
-     along any dimension in which chrominance subsampling is used), then an
-     intermediate buffer copy will be performed within TurboJPEG/OSS.
-  */
+  /* Nothing to see here.  Pay no attention to the man behind the curtain. */

 typedef void* tjhandle;

@@ -96,7 +77,8 @@ extern "C" {
  and returns a handle to the instance.  Most applications will only
  need to call this once at the beginning of the program or once for each
  concurrent thread.  Don't try to create a new instance every time you
-  compress an image, because this will cause performance to suffer.
+  compress an image, because this may cause performance to suffer in some
+  TurboJPEG implementations.

  RETURNS: NULL on error
 */
@@ -111,39 +93,40 @@ DLLEXPORT tjhandle DLLCALL tjInitCompress(void);

  [INPUT] j = instance handle previously returned from a call to
     tjInitCompress()
-  [INPUT] srcbuf = pointer to user-allocated image buffer containing pixels in
-     RGB(A) or BGR(A) form
-  [INPUT] width =  width (in pixels) of the source image
+  [INPUT] srcbuf = pointer to user-allocated image buffer containing RGB or
+     grayscale pixels to be compressed
+  [INPUT] width = width (in pixels) of the source image
  [INPUT] pitch = bytes per line of the source image (width*pixelsize if the
     bitmap is unpadded, else TJPAD(width*pixelsize) if each line of the bitmap
     is padded to the nearest 32-bit boundary, such as is the case for Windows
-     bitmaps.  You can also be clever and use this parameter to skip lines, etc.,
-     as long as the pitch is greater than 0.)
+     bitmaps.  You can also be clever and use this parameter to skip lines,
+     etc.  Setting this parameter to 0 is the equivalent of setting it to
+     width*pixelsize.
  [INPUT] height = height (in pixels) of the source image
  [INPUT] pixelsize = size (in bytes) of each pixel in the source image
-     RGBA and BGRA: 4, RGB and BGR: 3, Grayscale: 1
+     RGBX/BGRX/XRGB/XBGR: 4, RGB/BGR: 3, Grayscale: 1
  [INPUT] dstbuf = pointer to user-allocated image buffer which will receive
-     the JPEG image.  Use the macro TJBUFSIZE(width, height) to determine
+     the JPEG image.  Use the TJBUFSIZE(width, height) function to determine
     the appropriate size for this buffer based on the image width and height.
  [OUTPUT] size = pointer to unsigned long which receives the size (in bytes)
     of the compressed image
-  [INPUT] jpegsubsamp = Specifies either 4:2:0, 4:2:2, or 4:4:4 subsampling.
-     When the image is converted from the RGB to YCbCr colorspace as part of the
-     JPEG compression process, every other Cb and Cr (chrominance) pixel can be
-     discarded to produce a smaller image with little perceptible loss of
-     image clarity (the human eye is more sensitive to small changes in
-     brightness than small changes in color.)
+  [INPUT] jpegsubsamp = Specifies either 4:2:0, 4:2:2, 4:4:4, or grayscale
+     subsampling.  When the image is converted from the RGB to YCbCr colorspace
+     as part of the JPEG compression process, every other Cb and Cr
+     (chrominance) pixel can be discarded to produce a smaller image with
+     little perceptible loss of image clarity (the human eye is more sensitive
+     to small changes in brightness than small changes in color.)

     TJ_420: 4:2:0 subsampling.  Discards every other Cb, Cr pixel in both
-        horizontal and vertical directions.
+        horizontal and vertical directions
     TJ_422: 4:2:2 subsampling.  Discards every other Cb, Cr pixel only in
-        the horizontal direction.
-     TJ_444: no subsampling.
+        the horizontal direction
+     TJ_444: no subsampling
     TJ_GRAYSCALE: Generate grayscale JPEG image

-  [INPUT] jpegqual = JPEG quality (an integer between 0 and 100 inclusive.)
+  [INPUT] jpegqual = JPEG quality (an integer between 0 and 100 inclusive)
  [INPUT] flags = the bitwise OR of one or more of the flags described in the
-     "Flags" section above.
+     "Flags" section above

  RETURNS: 0 on success, -1 on error
 */
@@ -152,8 +135,77 @@ DLLEXPORT int DLLCALL tjCompress(tjhandle j,
 	unsigned char *dstbuf, unsigned long *size,
 	int jpegsubsamp, int jpegqual, int flags);

+
+/*
+  unsigned long TJBUFSIZE(int width, int height)
+
+  Convenience function which returns the maximum size of the buffer required to
+  hold a JPEG image with the given width and height
+
+  RETURNS: -1 if arguments are out of bounds
+*/
 DLLEXPORT unsigned long DLLCALL TJBUFSIZE(int width, int height);

+
+/*
+  unsigned long TJBUFSIZEYUV(int width, int height, int subsamp)
+
+  Convenience function which returns the size of the buffer required to
+  hold a YUV planar image with the given width, height, and level of
+  chrominance subsampling
+
+  RETURNS: -1 if arguments are out of bounds
+*/
+DLLEXPORT unsigned long DLLCALL TJBUFSIZEYUV(int width, int height,
+  int subsamp);
+
+
+/*
+  int tjEncodeYUV(tjhandle j,
+     unsigned char *srcbuf, int width, int pitch, int height, int pixelsize,
+     unsigned char *dstbuf, int subsamp, int flags)
+
+  This function uses the accelerated color conversion routines in TurboJPEG's
+  underlying codec to produce a planar YUV image that is suitable for X Video.
+  Specifically, if the chrominance components are subsampled along the
+  horizontal dimension, then the width of the luminance plane is padded to 2 in
+  the output image (same goes for the height of the luminance plane, if the
+  chrominance components are subsampled along the vertical dimension.)  Also,
+  each line of each plane in the output image is padded to 4 bytes.  Although
+  this will work with any subsampling option, it is really only useful in
+  combination with TJ_420, which produces an image compatible with the I420
+  (AKA "YUV420P") format.
+
+  [INPUT] j = instance handle previously returned from a call to
+     tjInitCompress()
+  [INPUT] srcbuf = pointer to user-allocated image buffer containing RGB or
+     grayscale pixels to be encoded
+  [INPUT] width = width (in pixels) of the source image
+  [INPUT] pitch = bytes per line of the source image (width*pixelsize if the
+     bitmap is unpadded, else TJPAD(width*pixelsize) if each line of the bitmap
+     is padded to the nearest 32-bit boundary, such as is the case for Windows
+     bitmaps.  You can also be clever and use this parameter to skip lines,
+     etc.  Setting this parameter to 0 is the equivalent of setting it to
+     width*pixelsize.
+  [INPUT] height = height (in pixels) of the source image
+  [INPUT] pixelsize = size (in bytes) of each pixel in the source image
+     RGBX/BGRX/XRGB/XBGR: 4, RGB/BGR: 3, Grayscale: 1
+  [INPUT] dstbuf = pointer to user-allocated image buffer which will receive
+     the YUV image.  Use the TJBUFSIZEYUV(width, height, subsamp) function to
+     determine the appropriate size for this buffer based on the image width,
+     height, and level of subsampling.
+  [INPUT] subsamp = Specifies either 4:2:0, 4:2:2, 4:4:4, or grayscale
+     subsampling (see description under tjCompress())
+  [INPUT] flags = the bitwise OR of one or more of the flags described in the
+     "Flags" section above
+
+  RETURNS: 0 on success, -1 on error
+*/
+DLLEXPORT int DLLCALL tjEncodeYUV(tjhandle j,
+	unsigned char *srcbuf, int width, int pitch, int height, int pixelsize,
+	unsigned char *dstbuf, int subsamp, int flags);
+
+
 /*
  tjhandle tjInitDecompress(void)

@@ -161,7 +213,8 @@ DLLEXPORT unsigned long DLLCALL TJBUFSIZE(int width, int height);
  structures, and returns a handle to the instance.  Most applications will
  only need to call this once at the beginning of the program or once for each
  concurrent thread.  Don't try to create a new instance every time you
-  decompress an image, because this will cause performance to suffer.
+  decompress an image, because this may cause performance to suffer in some
+  TurboJPEG implementations.

  RETURNS: NULL on error
 */
@@ -171,26 +224,25 @@ DLLEXPORT tjhandle DLLCALL tjInitDecompress(void);
 /*
  int tjDecompressHeader2(tjhandle j,
     unsigned char *srcbuf, unsigned long size,
-     int *width, int *height, int *jpegsub)
+     int *width, int *height, int *jpegsubsamp)

  [INPUT] j = instance handle previously returned from a call to
     tjInitDecompress()
-  [INPUT] srcbuf = pointer to a user-allocated buffer containing the JPEG image
-     to decompress
+  [INPUT] srcbuf = pointer to a user-allocated buffer containing a JPEG image
  [INPUT] size = size of the JPEG image buffer (in bytes)
  [OUTPUT] width = width (in pixels) of the JPEG image
  [OUTPUT] height = height (in pixels) of the JPEG image
-  [OUTPUT] jpegsub = type of chrominance subsampling used when compressing the
-     JPEG image
+  [OUTPUT] jpegsubsamp = type of chrominance subsampling used when compressing
+     the JPEG image

  RETURNS: 0 on success, -1 on error
 */
 DLLEXPORT int DLLCALL tjDecompressHeader2(tjhandle j,
 	unsigned char *srcbuf, unsigned long size,
-	int *width, int *height, int *jpegsub);
+	int *width, int *height, int *jpegsubsamp);

 /*
-  Deprecated version of the above function
+  Legacy version of the above function
 */
 DLLEXPORT int DLLCALL tjDecompressHeader(tjhandle j,
 	unsigned char *srcbuf, unsigned long size,
@@ -212,15 +264,16 @@ DLLEXPORT int DLLCALL tjDecompressHeader(tjhandle j,
     the bitmap image.  This buffer should normally be pitch*height
     bytes in size, although this pointer may also be used to decompress into
     a specific region of a larger buffer.
-  [INPUT] width =  width (in pixels) of the destination image
-  [INPUT] pitch = bytes per line of the destination image (width*pixelsize if the
-     bitmap is unpadded, else TJPAD(width*pixelsize) if each line of the bitmap
-     is padded to the nearest 32-bit boundary, such as is the case for Windows
-     bitmaps.  You can also be clever and use this parameter to skip lines, etc.,
-     as long as the pitch is greater than 0.)
+  [INPUT] width = width (in pixels) of the destination image
+  [INPUT] pitch = bytes per line of the destination image (width*pixelsize if
+     the bitmap is unpadded, else TJPAD(width*pixelsize) if each line of the
+     bitmap is padded to the nearest 32-bit boundary, such as is the case for
+     Windows bitmaps.  You can also be clever and use this parameter to skip
+     lines, etc.  Setting this parameter to 0 is the equivalent of setting it
+     to width*pixelsize.
  [INPUT] height = height (in pixels) of the destination image
  [INPUT] pixelsize = size (in bytes) of each pixel in the destination image
-     RGBA/RGBx and BGRA/BGRx: 4, RGB and BGR: 3, Grayscale: 1
+     RGBX/BGRX/XRGB/XBGR: 4, RGB/BGR: 3, Grayscale: 1
  [INPUT] flags = the bitwise OR of one or more of the flags described in the
     "Flags" section above.

@@ -232,6 +285,37 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle j,
 	int flags);


+/*
+  int tjDecompressToYUV(tjhandle j,
+     unsigned char *srcbuf, unsigned long size,
+     unsigned char *dstbuf, int flags)
+
+  This function performs JPEG decompression but leaves out the color conversion
+  step, so a planar YUV image is generated instead of an RGB image.  The
+  padding of the planes in this image is the same as in tjEncodeYUV().
+  Note that, if the width or height of the output image is not a multiple of 8
+  (or a multiple of 16 along any dimension in which chrominance subsampling is
+  used), then an intermediate buffer copy will be performed within TurboJPEG.
+
+  [INPUT] j = instance handle previously returned from a call to
+     tjInitDecompress()
+  [INPUT] srcbuf = pointer to a user-allocated buffer containing the JPEG image
+     to decompress
+  [INPUT] size = size of the JPEG image buffer (in bytes)
+  [INPUT] dstbuf = pointer to user-allocated image buffer which will receive
+     the YUV image.  Use the TJBUFSIZEYUV(width, height, subsamp) function to
+     determine the appropriate size for this buffer based on the image width,
+     height, and level of subsampling.
+  [INPUT] flags = the bitwise OR of one or more of the flags described in the
+     "Flags" section above.
+
+  RETURNS: 0 on success, -1 on error
+*/
+DLLEXPORT int DLLCALL tjDecompressToYUV(tjhandle j,
+	unsigned char *srcbuf, unsigned long size,
+	unsigned char *dstbuf, int flags);
+
+
 /*
  int tjDestroy(tjhandle h)

--- a/turbojpegl.c
+++ b/turbojpegl.c
@@ -1,6 +1,6 @@
 /* Copyright (C)2004 Landmark Graphics Corporation
 * Copyright (C)2005 Sun Microsystems, Inc.
- * Copyright (C)2009-2010 D. R. Commander
+ * Copyright (C)2009-2011 D. R. Commander
 *
 * This library is free software and may be redistributed and/or modified under
 * the terms of the wxWindows Library License, Version 3.1 or (at your option)
@@ -70,10 +70,9 @@ static const int hsampfactor[NUMSUBOPT]={1, 2, 2, 1};
 static const int vsampfactor[NUMSUBOPT]={1, 1, 2, 1};
 static const int pixelsize[NUMSUBOPT]={3, 3, 3, 1};

-#define _throw(c) {sprintf(lasterror, "%s", c);  return -1;}
-#define _catch(f) {if((f)==-1) return -1;}
+#define _throw(c) {sprintf(lasterror, "%s", c);  retval=-1;  goto bailout;}
 #define checkhandle(h) jpgstruct *j=(jpgstruct *)h; \
-	if(!j) _throw("Invalid handle");
+	if(!j) {sprintf(lasterror, "Invalid handle");  return -1;}


 // CO
@@ -113,18 +112,46 @@ DLLEXPORT tjhandle DLLCALL tjInitCompress(void)
 	return (tjhandle)j;
 }

+
 DLLEXPORT unsigned long DLLCALL TJBUFSIZE(int width, int height)
 {
-	// This allows enough room in case the image doesn't compress
-	return ((width+15)&(~15)) * ((height+15)&(~15)) * 6 + 2048;
+	unsigned long retval=0;
+	if(width<1 || height<1)
+		_throw("Invalid argument in TJBUFSIZE()");
+
+	// This allows for rare corner cases in which a JPEG image can actually be
+	// larger than the uncompressed input (we wouldn't mention it if it hadn't
+	// happened before.)
+	retval=((width+15)&(~15)) * ((height+15)&(~15)) * 6 + 2048;
+
+	bailout:
+	return retval;
 }

+
+DLLEXPORT unsigned long DLLCALL TJBUFSIZEYUV(int width, int height,
+	int subsamp)
+{
+	unsigned long retval=0;
+	int pw, ph, cw, ch;
+	if(width<1 || height<1 || subsamp<0 || subsamp>=NUMSUBOPT)
+		_throw("Invalid argument in TJBUFSIZEYUV()");
+	pw=PAD(width, hsampfactor[subsamp]);
+	ph=PAD(height, vsampfactor[subsamp]);
+	cw=pw/hsampfactor[subsamp];  ch=ph/vsampfactor[subsamp];
+	retval=PAD(pw, 4)*ph + (subsamp==TJ_GRAYSCALE? 0:PAD(cw, 4)*ch*2);
+
+	bailout:
+	return retval;
+}
+
+
 DLLEXPORT int DLLCALL tjCompress(tjhandle h,
 	unsigned char *srcbuf, int width, int pitch, int height, int ps,
 	unsigned char *dstbuf, unsigned long *size,
 	int jpegsub, int qual, int flags)
 {
-	int i;  JSAMPROW *row_pointer=NULL;
+	int i, retval=0;  JSAMPROW *row_pointer=NULL;
 	JSAMPLE *_tmpbuf[MAX_COMPONENTS], *_tmpbuf2[MAX_COMPONENTS];
 	JSAMPROW *tmpbuf[MAX_COMPONENTS], *tmpbuf2[MAX_COMPONENTS];
 	JSAMPROW *outbuf[MAX_COMPONENTS];
@@ -174,16 +201,8 @@ DLLEXPORT int DLLCALL tjCompress(tjhandle h,

 	if(setjmp(j->jerr.jb))
 	{  // this will execute if LIBJPEG has an error
-		if(row_pointer) free(row_pointer);
-		for(i=0; i<MAX_COMPONENTS; i++)
-		{
-			if(tmpbuf[i]!=NULL) free(tmpbuf[i]);
-			if(_tmpbuf[i]!=NULL) free(_tmpbuf[i]);
-			if(tmpbuf2[i]!=NULL) free(tmpbuf2[i]);
-			if(_tmpbuf2[i]!=NULL) free(_tmpbuf2[i]);
-			if(outbuf[i]!=NULL) free(outbuf[i]);
-		}
-		return -1;
+		retval=-1;
+		goto bailout;
 	}

 	jpeg_set_defaults(&j->cinfo);
@@ -193,7 +212,8 @@ DLLEXPORT int DLLCALL tjCompress(tjhandle h,
 		jpeg_set_colorspace(&j->cinfo, JCS_GRAYSCALE);
 	else
 		jpeg_set_colorspace(&j->cinfo, JCS_YCbCr);
-	j->cinfo.dct_method = JDCT_FASTEST;
+	if(qual>=96) j->cinfo.dct_method=JDCT_ISLOW;
+	else j->cinfo.dct_method=JDCT_FASTEST;

 	j->cinfo.comp_info[0].h_samp_factor=hsampfactor[jpegsub];
 	j->cinfo.comp_info[1].h_samp_factor=1;
@@ -301,6 +321,8 @@ DLLEXPORT int DLLCALL tjCompress(tjhandle h,
 		*size=TJBUFSIZE(j->cinfo.image_width, j->cinfo.image_height)
 			-(unsigned long)(j->jdms.free_in_buffer);

+	bailout:
+	if(j->cinfo.global_state>CSTATE_START) jpeg_abort_compress(&j->cinfo);
 	if(row_pointer) free(row_pointer);
 	for(i=0; i<MAX_COMPONENTS; i++)
 	{
@@ -310,7 +332,17 @@ DLLEXPORT int DLLCALL tjCompress(tjhandle h,
 		if(_tmpbuf2[i]!=NULL) free(_tmpbuf2[i]);
 		if(outbuf[i]!=NULL) free(outbuf[i]);
 	}
-	return 0;
+	return retval;
+}
+
+
+DLLEXPORT int DLLCALL tjEncodeYUV(tjhandle h,
+	unsigned char *srcbuf, int width, int pitch, int height, int ps,
+	unsigned char *dstbuf, int subsamp, int flags)
+{
+	unsigned long size;
+	return tjCompress(h, srcbuf, width, pitch, height, ps, dstbuf, &size,
+		subsamp, 0, flags|TJ_YUV);
 }


@@ -360,45 +392,16 @@ DLLEXPORT tjhandle DLLCALL tjInitDecompress(void)
 }


-DLLEXPORT int DLLCALL tjDecompressHeader(tjhandle h,
-	unsigned char *srcbuf, unsigned long size,
-	int *width, int *height)
-{
-	checkhandle(h);
-
-	if(srcbuf==NULL || size<=0 || width==NULL || height==NULL)
-		_throw("Invalid argument in tjDecompressHeader()");
-	if(!j->initd) _throw("Instance has not been initialized for decompression");
-
-	if(setjmp(j->jerr.jb))
-	{  // this will execute if LIBJPEG has an error
-		return -1;
-	}
-
-	j->jsms.bytes_in_buffer = size;
-	j->jsms.next_input_byte = srcbuf;
-
-	jpeg_read_header(&j->dinfo, TRUE);
-
-	*width=j->dinfo.image_width;  *height=j->dinfo.image_height;
-
-	jpeg_abort_decompress(&j->dinfo);
-
-	if(*width<1 || *height<1) _throw("Invalid data returned in header");
-	return 0;
-}
-
-
 DLLEXPORT int DLLCALL tjDecompressHeader2(tjhandle h,
 	unsigned char *srcbuf, unsigned long size,
 	int *width, int *height, int *jpegsub)
 {
-	int i, k;
+	int i, k, retval=0;

 	checkhandle(h);

 	if(srcbuf==NULL || size<=0 || width==NULL || height==NULL || jpegsub==NULL)
-		_throw("Invalid argument in tjDecompressHeader()");
+		_throw("Invalid argument in tjDecompressHeader2()");
 	if(!j->initd) _throw("Instance has not been initialized for decompression");

 	if(setjmp(j->jerr.jb))
@@ -439,7 +442,18 @@ DLLEXPORT int DLLCALL tjDecompressHeader2(tjhandle h,

 	if(*jpegsub<0) _throw("Could not determine subsampling type for JPEG image");
 	if(*width<1 || *height<1) _throw("Invalid data returned in header");
-	return 0;
+
+	bailout:
+	return retval;
+}
+
+
+DLLEXPORT int DLLCALL tjDecompressHeader(tjhandle h,
+	unsigned char *srcbuf, unsigned long size,
+	int *width, int *height)
+{
+	int jpegsub;
+	return tjDecompressHeader2(h, srcbuf, size, width, height, &jpegsub);
 }


@@ -448,7 +462,7 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 	unsigned char *dstbuf, int width, int pitch, int height, int ps,
 	int flags)
 {
-	int i, row;  JSAMPROW *row_pointer=NULL, *outbuf[MAX_COMPONENTS];
+	int i, row, retval=0;  JSAMPROW *row_pointer=NULL, *outbuf[MAX_COMPONENTS];
 	int cw[MAX_COMPONENTS], ch[MAX_COMPONENTS], iw[MAX_COMPONENTS],
 		tmpbufsize=0, usetmpbuf=0, th[MAX_COMPONENTS];
 	JSAMPLE *_tmpbuf=NULL;  JSAMPROW *tmpbuf[MAX_COMPONENTS];
@@ -475,14 +489,8 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,

 	if(setjmp(j->jerr.jb))
 	{  // this will execute if LIBJPEG has an error
-		for(i=0; i<MAX_COMPONENTS; i++)
-		{
-			if(tmpbuf[i]!=NULL) free(tmpbuf[i]);
-			if(outbuf[i]!=NULL) free(outbuf[i]);
-		}
-		if(_tmpbuf) free(_tmpbuf);
-		if(row_pointer) free(row_pointer);
-		return -1;
+		retval=-1;
+		goto bailout;
 	}

 	j->jsms.bytes_in_buffer = size;
@@ -501,15 +509,15 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 			int ih;
 			iw[i]=compptr->width_in_blocks*DCTSIZE;
 			ih=compptr->height_in_blocks*DCTSIZE;
-			cw[i]=PAD(width, dinfo->max_h_samp_factor)*compptr->h_samp_factor
-				/dinfo->max_h_samp_factor;
-			ch[i]=PAD(height, dinfo->max_v_samp_factor)*compptr->v_samp_factor
-				/dinfo->max_v_samp_factor;
+			cw[i]=PAD(dinfo->image_width, dinfo->max_h_samp_factor)
+				*compptr->h_samp_factor/dinfo->max_h_samp_factor;
+			ch[i]=PAD(dinfo->image_height, dinfo->max_v_samp_factor)
+				*compptr->v_samp_factor/dinfo->max_v_samp_factor;
 			if(iw[i]!=cw[i] || ih!=ch[i]) usetmpbuf=1;
 			th[i]=compptr->v_samp_factor*DCTSIZE;
 			tmpbufsize+=iw[i]*th[i];
 			if((outbuf[i]=(JSAMPROW *)malloc(sizeof(JSAMPROW)*ch[i]))==NULL)
-				_throw("Memory allocation failed in tjInitDecompress()");
+				_throw("Memory allocation failed in tjDecompress()");
 			for(row=0; row<ch[i]; row++)
 			{
 				outbuf[i][row]=ptr;
@@ -519,13 +527,12 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 		if(usetmpbuf)
 		{
 			if((_tmpbuf=(JSAMPLE *)malloc(sizeof(JSAMPLE)*tmpbufsize))==NULL)
-				_throw("Memory allocation failed in tjInitDecompress()");
+				_throw("Memory allocation failed in tjDecompress()");
 			ptr=_tmpbuf;
 			for(i=0; i<dinfo->num_components; i++)
 			{
-				jpeg_component_info *compptr=&dinfo->comp_info[i];
 				if((tmpbuf[i]=(JSAMPROW *)malloc(sizeof(JSAMPROW)*th[i]))==NULL)
-					_throw("Memory allocation failed in tjInitDecompress()");
+					_throw("Memory allocation failed in tjDecompress()");
 				for(row=0; row<th[i]; row++)
 				{
 					tmpbuf[i][row]=ptr;
@@ -537,7 +544,7 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 	else
 	{
 		if((row_pointer=(JSAMPROW *)malloc(sizeof(JSAMPROW)*height))==NULL)
-			_throw("Memory allocation failed in tjInitDecompress()");
+			_throw("Memory allocation failed in tjDecompress()");
 		for(i=0; i<height; i++)
 		{
 			if(flags&TJ_BOTTOMUP) row_pointer[i]= &dstbuf[(height-i-1)*pitch];
@@ -587,7 +594,6 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 				int j;
 				for(i=0; i<dinfo->num_components; i++)
 				{
-					jpeg_component_info *compptr=&dinfo->comp_info[i];
 					for(j=0; j<min(th[i], ch[i]-crow[i]); j++)
 					{
 						memcpy(outbuf[i][crow[i]+j], tmpbuf[i][j], cw[i]);
@@ -606,6 +612,8 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 	}
 	jpeg_finish_decompress(&j->dinfo);

+	bailout:
+	if(j->dinfo.global_state>DSTATE_START) jpeg_abort_decompress(&j->dinfo);
 	for(i=0; i<MAX_COMPONENTS; i++)
 	{
 		if(tmpbuf[i]) free(tmpbuf[i]);
@@ -613,7 +621,15 @@ DLLEXPORT int DLLCALL tjDecompress(tjhandle h,
 	}
 	if(_tmpbuf) free(_tmpbuf);
 	if(row_pointer) free(row_pointer);
-	return 0;
+	return retval;
+}
+
+
+DLLEXPORT int DLLCALL tjDecompressToYUV(tjhandle h,
+	unsigned char *srcbuf, unsigned long size,
+	unsigned char *dstbuf, int flags)
+{
+	return tjDecompress(h, srcbuf, size, dstbuf, 1, 0, 1, 3, flags|TJ_YUV);
 }
Author	SHA1	Message	Date
DRC	da58a91d1e	Make Visual C++ happy git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@465 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-27 00:07:51 +00:00
DRC	0769f505bf	Fix compiler warnings git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@461 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-26 21:08:39 +00:00
DRC	b4a50ceefb	Enable full GCC compiler warnings by default git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@460 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-26 21:07:50 +00:00
DRC	361a637174	Fix uninitialized variables git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@455 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-26 19:46:27 +00:00
DRC	3484e0b160	Add RGB-->Grayscale YUV decode test git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@453 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 06:21:15 +00:00
DRC	d5a69afa5a	Fix the buffer size regression test so that it generates the intended pixel patterns git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@450 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 02:13:07 +00:00
DRC	0745c0fb6a	Correct memory allocation failure messages git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@447 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 01:31:13 +00:00
DRC	94476c648b	Only access TJBUFSIZEYUV symbol if YUV encoding is enabled, and add 1 to the buffer size; This allows jpgtest to be used to benchmark the old TurboJPEG/IPP libraries. git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@446 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 01:14:46 +00:00
DRC	0cc1279e03	Use correct size for YUV buffers git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@445 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 00:54:29 +00:00
DRC	10e3193373	git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@441 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 00:14:12 +00:00
DRC	eeab6957dd	Replace the TJ_YUV flag with two new API functions, and add TJBUFSIZEYUV() from trunk git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@440 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-25 00:02:04 +00:00
DRC	5cb1b68d39	Tiling doesn't work with YUV encoding either git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@439 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-24 22:13:20 +00:00
DRC	0b11d1c815	Fix various CR/LF issues git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@438 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-24 21:43:35 +00:00
DRC	582a46a59e	Ensure that libjpeg state is reset if an error occurs git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@437 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-24 21:16:10 +00:00
DRC	6e51e19b1c	Use a proper version script for libturbojpeg, since we've added functions to it since 1.0 git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@436 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-24 19:36:02 +00:00
DRC	2ca3eb5660	Oops. dumpbuf() was displaying only red components. git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@429 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-23 01:40:44 +00:00
DRC	404e82e070	Test all YUV decode paths git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@427 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-22 19:59:54 +00:00
DRC	7d9d434826	"TurboJPEG/OSS"="TurboJPEG", since the TurboJPEG header is used with other codecs besides libjpeg-turbo git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@420 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-22 00:31:27 +00:00
DRC	b4b31403c7	RGBA=RGBX, BGRA=BGRX, ABGR=XBGR, ARGB=XRGB git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@412 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-19 17:31:24 +00:00
DRC	2a28599343	Implement "ultra quiet mode" for generating output that is easy to copy/paste into a spreadsheet git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@411 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-19 17:20:11 +00:00
DRC	3e87c8b0a4	Wordsmithing git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@410 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-19 16:42:14 +00:00
DRC	db8df01c85	git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@409 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-19 00:01:11 +00:00
DRC	240d82f8be	Include jpgtest in Unix/Linux distribution packages git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@408 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-19 00:00:34 +00:00
DRC	8569c2f5a6	Include libjpeg run-time programs and jpgtest in installer; Move documentation in SDK distribution packages into doc\ subdirectory git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@407 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 23:49:42 +00:00
DRC	5d6f858235	Disable CXX compiler check git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@406 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 23:08:58 +00:00
DRC	f3281e8275	Oops. We still had the description left over from VirtualGL git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@403 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 22:40:34 +00:00
DRC	0d2d965364	Wordsmithed the project description git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@402 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 22:29:45 +00:00
DRC	16e761e59c	Include high-quality test image in distribution git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@401 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 21:20:50 +00:00
DRC	69e15442f3	1.1.0 git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@400 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 21:16:56 +00:00
DRC	da509e32c6	New C version of jpgtest needs libm git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@399 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 21:16:04 +00:00
DRC	23ce65d25d	Updated (C) git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@392 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 07:02:32 +00:00
DRC	f38eee0739	The API/ABI should be described as "libjpeg v8", not "libjpeg v8b", since there are multiple libjpeg releases with the same interface git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@391 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 07:00:38 +00:00
DRC	c4ef01fc66	Add high-quality JPEG tests to CMake git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@388 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 05:06:58 +00:00
DRC	ab4db657db	Various documentation tweaks git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@387 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 04:55:08 +00:00
DRC	f0a5ba7542	"integer forward DCT", not "forward integer DCT" git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@382 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 03:21:06 +00:00
DRC	7c1df0a430	Performance pitfalls git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@380 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 02:45:24 +00:00
DRC	bb4465f392	git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@379 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 02:41:03 +00:00
DRC	72abc29722	The SIMD quantization algorithm does not produce correct results with the fast forward integer DCT and JPEG qualities >= 98, so for now, use the non-SIMD quantization function under those circumstances. git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@378 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-18 01:45:24 +00:00
DRC	fc1970acef	Fix typo that was causing visual artifacts in grayscale JPEGs git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@376 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-17 07:51:35 +00:00
DRC	dbf80031ae	Not using Intel codec anymore git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@371 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-16 01:39:29 +00:00
DRC	8254df1a06	Document that pitch=0 == pitch=width*pixelsize git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@366 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-15 07:57:48 +00:00
DRC	91e86ba6cf	Ensure that YUV temp buffers get properly freed if a failure occurs git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@364 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-15 05:24:08 +00:00
DRC	cfef7f1bd8	Ensure that the exit status of the program indicates failure if one of the pixel value checks fails git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@362 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-15 05:17:20 +00:00
DRC	b8cc52bf1a	Ported jpgtest.cxx to pure C to avoid the need for a C++ compiler git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@360 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-15 03:59:11 +00:00
DRC	e27124ab39	git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@359 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-08 07:14:33 +00:00
DRC	2e4d044cab	Ported jpgtest.cxx to pure C to avoid the need for a C++ compiler git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@355 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-08 01:18:37 +00:00
DRC	daae6d966a	Clarify that the C wrappers and headers fall under the same license as the rest of the SIMD code git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@334 632fc199-4ca6-4c93-a231-07263d6284db	2011-02-02 05:41:15 +00:00
DRC	4a1c2d9693	git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@329 632fc199-4ca6-4c93-a231-07263d6284db	2011-01-28 08:00:10 +00:00
DRC	c4bce4857e	git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@327 632fc199-4ca6-4c93-a231-07263d6284db	2011-01-28 07:51:25 +00:00
DRC	a4efd4eec7	Create 1.1.x branch git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.1.x@325 632fc199-4ca6-4c93-a231-07263d6284db	2011-01-28 06:06:34 +00:00
DRC	2d94e261ec	The fast integer DCT degrades for qualities > 95, so use the slow integer DCT instead to ensure that perceptually lossless performance is maintained. git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@320 632fc199-4ca6-4c93-a231-07263d6284db	2011-01-25 06:52:31 +00:00
DRC	01e6909180	Clarifications git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@319 632fc199-4ca6-4c93-a231-07263d6284db	2011-01-06 01:19:43 +00:00