From f04837052e35c6eedb4816f79673ab51a1a53d52 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Sun, 14 Dec 2025 18:38:11 +0200 Subject: [PATCH 01/20] reworked dependency system --- CMakeLists.txt | 10 +- external/astcenc.cmake | 123 + external/basisu.cmake | 69 + external/cxxopts.cmake | 29 + external/fmt.cmake | 29 + external/lodepng.cmake | 30 + external/zstd.cmake | 46 + .../basisu_c_binding/inc/basisu_c_binding.h | 2 +- lib/CMakeLists.txt | 152 +- lib/src/astc_codec.cpp | 2 +- lib/src/basis_encode.cpp | 4 +- lib/src/basis_transcode.cpp | 2 +- lib/src/miniz_wrapper.cpp | 2 +- other_include/zstd_errors.h | 93 - tests/ktxdiff/CMakeLists.txt | 2 +- tests/ktxdiff/ktxdiff_main.cpp | 2 +- tests/transcodetests/CMakeLists.txt | 2 +- tools/imageio/CMakeLists.txt | 23 +- tools/imageio/external/jpgd.cpp | 3230 ++++ tools/imageio/external/jpgd.h | 347 + tools/imageio/external/tinyexr.h | 13581 ++++++++++++++++ tools/imageio/image.hpp | 4 +- tools/imageio/jpg.imageio/jpginput.cc | 2 +- tools/ktx/CMakeLists.txt | 4 +- tools/ktx/command_extract.cpp | 6 +- tools/ktx/metrics_utils.h | 4 +- 26 files changed, 17532 insertions(+), 268 deletions(-) create mode 100644 external/astcenc.cmake create mode 100644 external/basisu.cmake create mode 100644 external/cxxopts.cmake create mode 100644 external/fmt.cmake create mode 100644 external/lodepng.cmake create mode 100644 external/zstd.cmake delete mode 100644 other_include/zstd_errors.h create mode 100644 tools/imageio/external/jpgd.cpp create mode 100644 tools/imageio/external/jpgd.h create mode 100644 tools/imageio/external/tinyexr.h diff --git a/CMakeLists.txt b/CMakeLists.txt index 0b4681920f..1ac3e3822a 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -382,7 +382,7 @@ if(EMSCRIPTEN AND KTX_FEATURE_JS) ) set_target_properties( msc_basis_transcoder_js PROPERTIES OUTPUT_NAME "msc_basis_transcoder") - target_link_libraries(msc_basis_transcoder_js basisu_encoder) + target_link_libraries(msc_basis_transcoder_js basisu::basisu_encoder) add_custom_command( TARGET msc_basis_transcoder_js @@ -459,14 +459,10 @@ add_subdirectory(interface/basisu_c_binding) # except for building the ktx library. if((KTX_FEATURE_TOOLS OR KTX_FEATURE_TESTS) AND NOT TARGET fmt::fmt) # Always build fmt static - set(BUILD_SHARED_LIBS OFF) - set(FMT_INSTALL OFF) - set(FMT_SYSTEM_HEADERS ON) - add_subdirectory(external/fmt) - set(BUILD_SHARED_LIBS ${BUILD_SHARED_LIBS_RESET}) + include("external/fmt.cmake") endif() if(KTX_FEATURE_TOOLS AND NOT TARGET cxxopts::cxxopts) - add_subdirectory(external/cxxopts) + include("external/cxxopts.cmake") endif() # Tools diff --git a/external/astcenc.cmake b/external/astcenc.cmake new file mode 100644 index 0000000000..10165cfef3 --- /dev/null +++ b/external/astcenc.cmake @@ -0,0 +1,123 @@ +#################################################### +# ASTC Encoder. +#################################################### +include(FetchContent) +include(${KTX_ROOT_DIR}/cmake/compiler_query_genexs.cmake) + +if (TARGET astc::astcenc) + message(STATUS "Using prebuilt astc") + return() +endif() + +set(ASTC_VERSION 5.3.0) +message(STATUS "ASTC VERSION: ${ASTC_VERSION}") + + +# Determine most of the ASTC-related settings automatically. + +# On Linux and Windows only one architecture is supported at once. On +# macOS simultaneous multiple architectures are supported by the ASTC +# encoder but not by the BasisU encoder. The latter supports only SSE +# SIMD that is enabled by compile time defines which makes it not amenable +# to a standard Xcode universal binary build. To ensure BASISU_SUPPORT_SSE +# is disabled when building for multiple architectures use only the +# value of CMAKE_OSX_ARCHITECTURES to decide if a universal build +# has been requested. Do not expose the astc-encoder's +# ASTCENC_UNIVERSAL_BUILD configuration option. + +set(universal_build OFF) +if("${CMAKE_SYSTEM_NAME}" STREQUAL "Darwin") + # Check CMAKE_OSX_ARCHITECTURES for multiple architectures. + list(FIND CMAKE_OSX_ARCHITECTURES "$(ARCHS_STANDARD)" archs_standard) + list(LENGTH CMAKE_OSX_ARCHITECTURES architecture_count) + if(NOT ${archs_standard} EQUAL -1 OR architecture_count GREATER 1) + set(universal_build ON) + endif() + # Set ordinary variable to override astc-encoder option's ON default + # and hide the option. + set(ASTCENC_UNIVERSAL_BUILD ${universal_build}) +endif() + +# If we detect the user is doing a universal build defer to astc-encoder +# to pick the architectures. If a universal build has not been requested, +# present options to the user to turn off SIMD and, if running on x86_64 +# to choose which SIMD ISA to use. On arm64 always use Neon. On unknown +# CPUs use None. +# +# On x86_64, if neither ASTCENC_ISA_SSE41 nor ASTCENC_ISA_SSE2 are defined, +# choose ASTCENC_ISA_AVX2. If ASTCENC_ISA_AVX2 fails to compile user must +# choose another x86_64 option. + +if(NOT ${universal_build}) + set(ASTCENC_ISA_NATIVE OFF) + set(ASTCENC_ISA_NEON OFF) + if(NOT CPU_ARCHITECTURE STREQUAL x86_64) + set(ASTCENC_ISA_AVX2 OFF) + set(ASTCENC_ISA_SSE41 OFF) + set(ASTCENC_ISA_SSE2 OFF) + endif() + + # "" in the CACHE sets causes use of the existing documentation string. + if(${ASTCENC_ISA_NONE}) + set(ASTCENC_LIB_TARGET astcenc-none-static) + if(CPU_ARCHITECTURE STREQUAL x86_64) + set(ASTCENC_ISA_AVX2 OFF CACHE BOOL "" FORCE) + set(ASTCENC_ISA_SSE41 OFF CACHE BOOL "" FORCE) + set(ASTCENC_ISA_SSE2 OFF CACHE BOOL "" FORCE) + endif() + else() + if(CPU_ARCHITECTURE STREQUAL x86_64) + if (${ASTCENC_ISA_SSE41}) + set(ASTCENC_LIB_TARGET astcenc-sse4.1-static) + set(ASTCENC_ISA_AVX2 OFF CACHE BOOL "" FORCE) + set(ASTCENC_ISA_SSE2 OFF CACHE BOOL "" FORCE) + elseif (${ASTCENC_ISA_SSE2}) + set(ASTCENC_LIB_TARGET astcenc-sse2-static) + set(ASTCENC_ISA_AVX2 OFF CACHE BOOL "" FORCE) + set(ASTCENC_ISA_SSE41 OFF CACHE BOOL "" FORCE) + else() + set(ASTCENC_LIB_TARGET astcenc-avx2-static) + set(ASTCENC_ISA_AVX2 ON CACHE BOOL "" FORCE) + set(ASTCENC_ISA_SSE41 OFF CACHE BOOL "" FORCE) + set(ASTCENC_ISA_SSE2 OFF CACHE BOOL "" FORCE) + endif() + elseif(CPU_ARCHITECTURE STREQUAL armv8 OR CPU_ARCHITECTURE STREQUAL arm64) + set(ASTCENC_LIB_TARGET astcenc-neon-static) + set(ASTCENC_ISA_NEON ON) + set(ASTCENC_ISA_NONE OFF CACHE BOOL "" FORCE) + else() + message(STATUS "Unsupported ISA for ASTC on ${CPU_ARCHITECTURE} arch, using ASTCENC_ISA_NONE.") + set(ASTCENC_ISA_NONE ON CACHE BOOL "" FORCE) + set(ASTCENC_LIB_TARGET astcenc-none-static) + endif() + endif() +endif() + +# setting ASTCENC_DECOMPRESSOR to ON breaks the build, so force it to OFF +# and hide it from cmake-gui (by using type INTERNAL) +set(ASTCENC_DECOMPRESSOR OFF CACHE INTERNAL "") +set(ASTCENC_CLI OFF) # Only build as library not the CLI astcencoder +# Force static build for astc-encoder +set(BUILD_SHARED_LIBS OFF) + +# Declare package +FetchContent_Declare( + astcenc + GIT_REPOSITORY https://github.com/ARM-software/astc-encoder + GIT_TAG ${ASTC_VERSION} + FIND_PACKAGE_ARGS +) + +# Populate astc +FetchContent_MakeAvailable(astcenc) + +set(BUILD_SHARED_LIBS ${BUILD_SHARED_LIBS_RESET}) +set_property(TARGET ${ASTCENC_LIB_TARGET} PROPERTY POSITION_INDEPENDENT_CODE ON) + +target_compile_definitions( + ${ASTCENC_LIB_TARGET} PRIVATE + $<$,19.40.33811>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> + $<$,17.0.3>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> +) + +add_library(astc::astcenc ALIAS ${ASTCENC_LIB_TARGET}) diff --git a/external/basisu.cmake b/external/basisu.cmake new file mode 100644 index 0000000000..682357a836 --- /dev/null +++ b/external/basisu.cmake @@ -0,0 +1,69 @@ +#################################################### +# Basis Universal Encoder. +#################################################### +include(FetchContent) + +if (TARGET basisu::basisu_encoder) + message(STATUS "Using prebuilt basisu") + return() +endif() + +#set(BASISU_VERSION 1_60) +#message(STATUS "Basisu VERSION: ${BASISU_VERSION}") + +# Options +set(TOOL FALSE) +set(EXAMPLES FALSE) +if(NOT ${CPU_ARCHITECTURE} STREQUAL "x86_64") + # Basisu sets this TRUE if MSVC is TRUE. + set(BASISU_SSE FALSE) +endif() + +# Declare package +FetchContent_Declare( + basisu + GIT_REPOSITORY https://github.com/Daniil-SV/basis_universal.git + GIT_TAG master #"v${BASISU_VERSION}" + #FIND_PACKAGE_ARGS +) + +# Populate basisu +FetchContent_MakeAvailable(basisu) + +if (NOT TARGET basisu::basisu_encoder) + add_library(basisu::basisu_encoder ALIAS basisu_encoder) + + # In some package managers, like vcpkg, all headers are stored in "basisu" subdirectory, + # while in the official repository they are stored in the root folder, which creates some conflicts. + # Objectively, headers in a subdirectory look safer and better. + # So the solution is to copy the headers to the subdirectory by hands, if necessary. + + # Temporary folder with our headers + set(BASISU_INCLUDE_DIR ${CMAKE_CURRENT_BINARY_DIR}/include/basisu) + + # Folders with headers in repo root + set(BASISU_SUBDIRS encoder transcoder) + foreach(subdir ${BASISU_SUBDIRS}) + # Collecting just headers paths + file(GLOB_RECURSE headers + RELATIVE ${basisu_SOURCE_DIR}/${subdir} + ${basisu_SOURCE_DIR}/${subdir}/*.h + ${basisu_SOURCE_DIR}/${subdir}/*.inc + ) + + # Iterating over collected files + foreach(filename ${headers}) + get_filename_component(directory ${filename} DIRECTORY) + file(${BASISU_INCLUDE_DIR} ${BASISU_INCLUDE_DIR}/${subdir}/${directory}) + + # And finally copy to temporary folder + file(COPY ${basisu_SOURCE_DIR}/${subdir}/${filename} + DESTINATION ${BASISU_INCLUDE_DIR}/${subdir}/${directory}) + endforeach() + endforeach() + + # Add temporary include path to target + target_include_directories(basisu_encoder + PUBLIC ${CMAKE_CURRENT_BINARY_DIR}/include + ) +endif() diff --git a/external/cxxopts.cmake b/external/cxxopts.cmake new file mode 100644 index 0000000000..c4a90a48ac --- /dev/null +++ b/external/cxxopts.cmake @@ -0,0 +1,29 @@ +#################################################### +# cxxopts +#################################################### +include(FetchContent) + +if (TARGET cxxopts::cxxopts) + message(STATUS "Using prebuilt cxxopts") + return() +endif() + +set(CXXOPTS_VERSION 3.3.1) +message(STATUS "CXXOPTS VERSION: ${CXXOPTS_VERSION}") + +# Options +# Always build fmt static +set(BUILD_SHARED_LIBS OFF) +set(FMT_INSTALL OFF) +set(FMT_SYSTEM_HEADERS ON) + +# Declare package +FetchContent_Declare( + cxxopts + GIT_REPOSITORY https://github.com/jarro2783/cxxopts.git + GIT_TAG "v${CXXOPTS_VERSION}" + FIND_PACKAGE_ARGS +) + +# Populate fmt +FetchContent_MakeAvailable(cxxopts) \ No newline at end of file diff --git a/external/fmt.cmake b/external/fmt.cmake new file mode 100644 index 0000000000..b83aef008f --- /dev/null +++ b/external/fmt.cmake @@ -0,0 +1,29 @@ +#################################################### +# fmt +#################################################### +include(FetchContent) + +if (TARGET fmt::fmt) + message(STATUS "Using prebuilt fmt") + return() +endif() + +set(FMT_VERSION 11.2.0) +message(STATUS "FMT VERSION: ${FMT_VERSION}") + +# Options +# Always build fmt static +set(BUILD_SHARED_LIBS OFF) +set(FMT_INSTALL OFF) +set(FMT_SYSTEM_HEADERS ON) + +# Declare package +FetchContent_Declare( + fmt + GIT_REPOSITORY https://github.com/fmtlib/fmt.git + GIT_TAG "v${FMT_VERSION}" + FIND_PACKAGE_ARGS +) + +# Populate fmt +FetchContent_MakeAvailable(fmt) diff --git a/external/lodepng.cmake b/external/lodepng.cmake new file mode 100644 index 0000000000..6f9dab0c24 --- /dev/null +++ b/external/lodepng.cmake @@ -0,0 +1,30 @@ +#################################################### +# lodepng +#################################################### +include(FetchContent) + +if (TARGET lodepng::lodepng) + message(STATUS "Using prebuilt lodepng") + return() +endif() + +# Declare package +FetchContent_Declare( + lodepng + GIT_REPOSITORY https://github.com/KhronosGroup/lodepng.git + GIT_TAG d4cc52d8c074da137303e1117cc098a97647960c +) + +# Populate lodepng +FetchContent_MakeAvailable(lodepng) + +add_library( + lodepng STATIC + ${lodepng_SOURCE_DIR}/lodepng.cpp + ${lodepng_SOURCE_DIR}/lodepng.h +) +add_library(lodepng::lodepng ALIAS lodepng) +target_include_directories( + lodepng PUBLIC + ${lodepng_SOURCE_DIR} +) diff --git a/external/zstd.cmake b/external/zstd.cmake new file mode 100644 index 0000000000..991bb01028 --- /dev/null +++ b/external/zstd.cmake @@ -0,0 +1,46 @@ +#################################################### +# Zstandart +#################################################### +include(FetchContent) + +if (TARGET zstd::libzstd) + message(STATUS "Using prebuilt zstd") + return() +endif() + +set(ZSTD_VERSION 1.5.7) +message(STATUS "ZSTD VERSION: ${ZSTD_VERSION}") + +# Options +# Build static by default +if(NOT DEFINED ZSTD_BUILD_SHARED) + set(ZSTD_BUILD_STATIC ON) + set(ZSTD_BUILD_SHARED OFF) +endif() +set(ZSTD_LEGACY_SUPPORT OFF) +set(ZSTD_BUILD_DICTBUILDER OFF) +set(ZSTD_BUILD_DEPRECATED OFF) +set(ZSTD_BUILD_PROGRAMS OFF) +set(ZSTD_BUILD_TESTS OFF) +set(ZSTD_BUILD_CONTRIB OFF) + +# Declare package +FetchContent_Declare( + zstd + DOWNLOAD_EXTRACT_TIMESTAMP TRUE + SOURCE_SUBDIR build/cmake + URL "https://github.com/facebook/zstd/releases/download/v${ZSTD_VERSION}/zstd-${ZSTD_VERSION}.tar.gz" + FIND_PACKAGE_ARGS NAMES ZSTD +) + +# Populate zstd +FetchContent_MakeAvailable(zstd) + +if (NOT TARGET zstd::libzstd) + # Create one common target name + if (TARGET libzstd_static) + add_library(zstd::libzstd ALIAS libzstd_static) + elseif(TARGET libzstd_shared) + add_library(zstd::libzstd ALIAS libzstd_shared) + endif() +endif() \ No newline at end of file diff --git a/interface/basisu_c_binding/inc/basisu_c_binding.h b/interface/basisu_c_binding/inc/basisu_c_binding.h index 02061cbf47..f06260ba55 100644 --- a/interface/basisu_c_binding/inc/basisu_c_binding.h +++ b/interface/basisu_c_binding/inc/basisu_c_binding.h @@ -18,7 +18,7 @@ #if defined(__GNUC__) && !defined(__clang__) #pragma GCC diagnostic ignored "-Wunused-value" #endif -#include +#include "basisu/transcoder/basisu_transcoder.h" #if defined(__GCC__) && !defined(__clang__) #pragma GCC diagnostic pop #endif diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index 0e6f165b89..bb48786135 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -128,6 +128,11 @@ elseif(LINUX) find_package(Threads REQUIRED) endif() +# Dependencies +include("../external/zstd.cmake") +include("../external/astcenc.cmake") +include("../external/basisu.cmake") + # Source Lists set(LIBKTX_MAIN_SRC @@ -467,6 +472,16 @@ macro(common_libktx_settings target enable_write library_type) $<$:${OpenCL_LIBRARY}> ) endif() + + target_link_libraries( + ${target} PRIVATE + zstd::libzstd + ) + + target_link_libraries( + ${target} PRIVATE + astc::astcenc + ) endmacro(common_libktx_settings) macro(add_lib_dependencies target lib) @@ -479,19 +494,6 @@ macro(add_lib_dependencies target lib) endif() endmacro(add_lib_dependencies) -#################################################### -# Basis Universal Encoder. -#################################################### - -set(BASISU_TOOL FALSE) -set(BASISU_EXAMPLES FALSE) -if(NOT ${CPU_ARCHITECTURE} STREQUAL "x86_64") - # Basisu sets this TRUE if MSVC is TRUE. - set(BASISU_SSE FALSE) -endif() -#set(BASISU_ZSTD FALSE) -add_subdirectory(../external/basis_universal basisu-encoder) - ################################################### # ktx libraries ################################################### @@ -504,7 +506,7 @@ if(LIBKTX_VERSION_FULL) src/writer2.c ) common_libktx_settings(ktx 1 ${LIB_TYPE}) - add_lib_dependencies(ktx basisu_encoder) + add_lib_dependencies(ktx basisu::basisu_encoder) endif() # Read-only library @@ -513,7 +515,7 @@ if(LIBKTX_VERSION_READ_ONLY) ${LIBKTX_MAIN_SRC} ) common_libktx_settings(ktx_read 0 ${LIB_TYPE}) - add_lib_dependencies(ktx_read basisu_encoder) + add_lib_dependencies(ktx_read basisu::basisu_encoder) endif() create_version_header(src ktx) @@ -545,124 +547,6 @@ if(NOT BUILD_SHARED_LIBS AND ) endif() - -#################################################### -# ASTC Encoder. -#################################################### - -# Determine most of the ASTC-related settings automatically. - -# On Linux and Windows only one architecture is supported at once. On -# macOS simultaneous multiple architectures are supported by the ASTC -# encoder but not by the BasisU encoder. The latter supports only SSE -# SIMD that is enabled by compile time defines which makes it not amenable -# to a standard Xcode universal binary build. To ensure BASISU_SUPPORT_SSE -# is disabled when building for multiple architectures use only the -# value of CMAKE_OSX_ARCHITECTURES to decide if a universal build -# has been requested. Do not expose the astc-encoder's -# ASTCENC_UNIVERSAL_BUILD configuration option. - -set(universal_build OFF) -if("${CMAKE_SYSTEM_NAME}" STREQUAL "Darwin") - # Check CMAKE_OSX_ARCHITECTURES for multiple architectures. - list(FIND CMAKE_OSX_ARCHITECTURES "$(ARCHS_STANDARD)" archs_standard) - list(LENGTH CMAKE_OSX_ARCHITECTURES architecture_count) - if(NOT ${archs_standard} EQUAL -1 OR architecture_count GREATER 1) - set(universal_build ON) - endif() - # Set ordinary variable to override astc-encoder option's ON default - # and hide the option. - set(ASTCENC_UNIVERSAL_BUILD ${universal_build}) -endif() - -# If we detect the user is doing a universal build defer to astc-encoder -# to pick the architectures. If a universal build has not been requested, -# present options to the user to turn off SIMD and, if running on x86_64 -# to choose which SIMD ISA to use. On arm64 always use Neon. On unknown -# CPUs use None. -# -# On x86_64, if neither ASTCENC_ISA_SSE41 nor ASTCENC_ISA_SSE2 are defined, -# choose ASTCENC_ISA_AVX2. If ASTCENC_ISA_AVX2 fails to compile user must -# choose another x86_64 option. - -if(NOT ${universal_build}) - set(ASTCENC_ISA_NATIVE OFF) - set(ASTCENC_ISA_NEON OFF) - if(NOT CPU_ARCHITECTURE STREQUAL x86_64) - set(ASTCENC_ISA_AVX2 OFF) - set(ASTCENC_ISA_SSE41 OFF) - set(ASTCENC_ISA_SSE2 OFF) - endif() - - # "" in the CACHE sets causes use of the existing documentation string. - if(${ASTCENC_ISA_NONE}) - set(ASTCENC_LIB_TARGET astcenc-none-static) - if(CPU_ARCHITECTURE STREQUAL x86_64) - set(ASTCENC_ISA_AVX2 OFF CACHE BOOL "" FORCE) - set(ASTCENC_ISA_SSE41 OFF CACHE BOOL "" FORCE) - set(ASTCENC_ISA_SSE2 OFF CACHE BOOL "" FORCE) - endif() - else() - if(CPU_ARCHITECTURE STREQUAL x86_64) - if (${ASTCENC_ISA_SSE41}) - set(ASTCENC_LIB_TARGET astcenc-sse4.1-static) - set(ASTCENC_ISA_AVX2 OFF CACHE BOOL "" FORCE) - set(ASTCENC_ISA_SSE2 OFF CACHE BOOL "" FORCE) - elseif (${ASTCENC_ISA_SSE2}) - set(ASTCENC_LIB_TARGET astcenc-sse2-static) - set(ASTCENC_ISA_AVX2 OFF CACHE BOOL "" FORCE) - set(ASTCENC_ISA_SSE41 OFF CACHE BOOL "" FORCE) - else() - set(ASTCENC_LIB_TARGET astcenc-avx2-static) - set(ASTCENC_ISA_AVX2 ON CACHE BOOL "" FORCE) - set(ASTCENC_ISA_SSE41 OFF CACHE BOOL "" FORCE) - set(ASTCENC_ISA_SSE2 OFF CACHE BOOL "" FORCE) - endif() - elseif(CPU_ARCHITECTURE STREQUAL armv8 OR CPU_ARCHITECTURE STREQUAL arm64) - set(ASTCENC_LIB_TARGET astcenc-neon-static) - set(ASTCENC_ISA_NEON ON) - set(ASTCENC_ISA_NONE OFF CACHE BOOL "" FORCE) - else() - message(STATUS "Unsupported ISA for ASTC on ${CPU_ARCHITECTURE} arch, using ASTCENC_ISA_NONE.") - set(ASTCENC_ISA_NONE ON CACHE BOOL "" FORCE) - set(ASTCENC_LIB_TARGET astcenc-none-static) - endif() - endif() -endif() - -# setting ASTCENC_DECOMPRESSOR to ON breaks the build, so force it to OFF -# and hide it from cmake-gui (by using type INTERNAL) -set(ASTCENC_DECOMPRESSOR OFF CACHE INTERNAL "") -set(ASTCENC_CLI OFF) # Only build as library not the CLI astcencoder -# Force static build for astc-encoder -set(BUILD_SHARED_LIBS OFF) -add_subdirectory(../external/astc-encoder astc-encoder) -set(BUILD_SHARED_LIBS ${BUILD_SHARED_LIBS_RESET}) -set_property(TARGET ${ASTCENC_LIB_TARGET} PROPERTY POSITION_INDEPENDENT_CODE ON) - -target_compile_definitions( - ${ASTCENC_LIB_TARGET} -PRIVATE - # ASTC encoder uses std::mutex. For more info. see comment about same - # setting in common_libktx_settings starting about line 355. To be eventually - # removed as noted in that comment. - $<$,19.40.33811>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> - $<$,17.0.3>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> -) - -if(NOT PROJECT_IS_TOP_LEVEL AND ${CMAKE_PROJECT_NAME} STREQUAL "KTX-Software") - # ktxdiff needs access to this variable to find the ASTC - # library to link with so set it in parent. - set(ASTCENC_LIB_TARGET ${ASTCENC_LIB_TARGET} PARENT_SCOPE) -endif() - -if(LIBKTX_VERSION_FULL) - add_lib_dependencies(ktx ${ASTCENC_LIB_TARGET}) -endif() -if(LIBKTX_VERSION_READ_ONLY) - add_lib_dependencies(ktx_read ${ASTCENC_LIB_TARGET}) -endif() - # Install if(LIBKTX_VERSION_FULL) @@ -675,7 +559,7 @@ if(LIBKTX_VERSION_READ_ONLY AND NOT (BUILD_SHARED_LIBRARIES OR LIBKTX_VERSION_FU list(APPEND LIBKTX_INSTALL_TARGETS ktx_read) endif() if(NOT BUILD_SHARED_LIBS AND NOT APPLE) - list(APPEND LIBKTX_INSTALL_TARGETS ${ASTCENC_LIB_TARGET} basisu_encoder) + list(APPEND LIBKTX_INSTALL_TARGETS astc::astcenc basisu::basisu_encoder) endif() if(APPLE OR LINUX) diff --git a/lib/src/astc_codec.cpp b/lib/src/astc_codec.cpp index afecbb6e3c..e22957b4fd 100644 --- a/lib/src/astc_codec.cpp +++ b/lib/src/astc_codec.cpp @@ -32,7 +32,7 @@ #include "texture2.h" #include "vkformat_enum.h" -#include "astc-encoder/Source/astcenc.h" +#include //************************************************************************ //* Functions common to decoder and encoder * diff --git a/lib/src/basis_encode.cpp b/lib/src/basis_encode.cpp index 3d20b63256..6621f89ae3 100644 --- a/lib/src/basis_encode.cpp +++ b/lib/src/basis_encode.cpp @@ -37,8 +37,8 @@ // If Rich does not accept the warning fixes need to add -Wunused-local-typedef for clang here. #pragma GCC diagnostic ignored "-Wunused-value" #endif -#include "encoder/basisu_comp.h" -#include "transcoder/basisu_transcoder.h" +#include "basisu/encoder/basisu_comp.h" +#include "basisu/transcoder/basisu_transcoder.h" #if defined(__GCC__) && !defined(__clang__) #pragma GCC diagnostic pop #endif diff --git a/lib/src/basis_transcode.cpp b/lib/src/basis_transcode.cpp index 4b1d86c38d..f7a36d9216 100644 --- a/lib/src/basis_transcode.cpp +++ b/lib/src/basis_transcode.cpp @@ -33,7 +33,7 @@ // If Rich does not accept the warning fixes need to add -Wunused-local-typedef for clang here. #pragma GCC diagnostic ignored "-Wunused-value" #endif -#include "transcoder/basisu_transcoder.h" +#include "basisu/transcoder/basisu_transcoder.h" #if defined(__GCC__) && !defined(__clang__) #pragma GCC diagnostic pop #endif diff --git a/lib/src/miniz_wrapper.cpp b/lib/src/miniz_wrapper.cpp index bbbd9a802b..9fe558a66a 100644 --- a/lib/src/miniz_wrapper.cpp +++ b/lib/src/miniz_wrapper.cpp @@ -30,7 +30,7 @@ #pragma GCC diagnostic ignored "-Wextra" #pragma GCC diagnostic ignored "-Wmisleading-indentation" #endif -#include "encoder/basisu_miniz.h" +#include "basisu/encoder/basisu_miniz.h" #ifdef __GNUC__ #pragma GCC diagnostic pop #endif diff --git a/other_include/zstd_errors.h b/other_include/zstd_errors.h deleted file mode 100644 index 238309457b..0000000000 --- a/other_include/zstd_errors.h +++ /dev/null @@ -1,93 +0,0 @@ -/* - * Copyright (c) 2016-2020, Yann Collet, Facebook, Inc. - * All rights reserved. - * - * This source code is licensed under both the BSD-style license (found in the - * LICENSE file in the root directory of this source tree) and the GPLv2 (found - * in the COPYING file in the root directory of this source tree). - * You may select, at your option, one of the above-listed licenses. - */ - -#ifndef ZSTD_ERRORS_H_398273423 -#define ZSTD_ERRORS_H_398273423 - -#if defined (__cplusplus) -extern "C" { -#endif - -/*===== dependency =====*/ -#include /* size_t */ - - -/* ===== ZSTDERRORLIB_API : control library symbols visibility ===== */ -#ifndef ZSTDERRORLIB_VISIBILITY -# if defined(__GNUC__) && (__GNUC__ >= 4) -# define ZSTDERRORLIB_VISIBILITY __attribute__ ((visibility ("default"))) -# else -# define ZSTDERRORLIB_VISIBILITY -# endif -#endif -#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1) -# define ZSTDERRORLIB_API __declspec(dllexport) ZSTDERRORLIB_VISIBILITY -#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1) -# define ZSTDERRORLIB_API __declspec(dllimport) ZSTDERRORLIB_VISIBILITY /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/ -#else -# define ZSTDERRORLIB_API ZSTDERRORLIB_VISIBILITY -#endif - -/*-********************************************* - * Error codes list - *-********************************************* - * Error codes _values_ are pinned down since v1.3.1 only. - * Therefore, don't rely on values if you may link to any version < v1.3.1. - * - * Only values < 100 are considered stable. - * - * note 1 : this API shall be used with static linking only. - * dynamic linking is not yet officially supported. - * note 2 : Prefer relying on the enum than on its value whenever possible - * This is the only supported way to use the error list < v1.3.1 - * note 3 : ZSTD_isError() is always correct, whatever the library version. - **********************************************/ -typedef enum { - ZSTD_error_no_error = 0, - ZSTD_error_GENERIC = 1, - ZSTD_error_prefix_unknown = 10, - ZSTD_error_version_unsupported = 12, - ZSTD_error_frameParameter_unsupported = 14, - ZSTD_error_frameParameter_windowTooLarge = 16, - ZSTD_error_corruption_detected = 20, - ZSTD_error_checksum_wrong = 22, - ZSTD_error_dictionary_corrupted = 30, - ZSTD_error_dictionary_wrong = 32, - ZSTD_error_dictionaryCreation_failed = 34, - ZSTD_error_parameter_unsupported = 40, - ZSTD_error_parameter_outOfBound = 42, - ZSTD_error_tableLog_tooLarge = 44, - ZSTD_error_maxSymbolValue_tooLarge = 46, - ZSTD_error_maxSymbolValue_tooSmall = 48, - ZSTD_error_stage_wrong = 60, - ZSTD_error_init_missing = 62, - ZSTD_error_memory_allocation = 64, - ZSTD_error_workSpace_tooSmall= 66, - ZSTD_error_dstSize_tooSmall = 70, - ZSTD_error_srcSize_wrong = 72, - ZSTD_error_dstBuffer_null = 74, - /* following error codes are __NOT STABLE__, they can be removed or changed in future versions */ - ZSTD_error_frameIndex_tooLarge = 100, - ZSTD_error_seekableIO = 102, - ZSTD_error_maxCode = 120 /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */ -} ZSTD_ErrorCode; - -/*! ZSTD_getErrorCode() : - convert a `size_t` function result into a `ZSTD_ErrorCode` enum type, - which can be used to compare with enum list published above */ -ZSTDERRORLIB_API ZSTD_ErrorCode ZSTD_getErrorCode(size_t functionResult); -ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code); /**< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */ - - -#if defined (__cplusplus) -} -#endif - -#endif /* ZSTD_ERRORS_H_398273423 */ diff --git a/tests/ktxdiff/CMakeLists.txt b/tests/ktxdiff/CMakeLists.txt index ba787e8ee5..d5e978702b 100644 --- a/tests/ktxdiff/CMakeLists.txt +++ b/tests/ktxdiff/CMakeLists.txt @@ -33,7 +33,7 @@ target_link_libraries( ktxdiff PRIVATE ktx - ${ASTCENC_LIB_TARGET} + astc::astcenc fmt::fmt objUtil ) diff --git a/tests/ktxdiff/ktxdiff_main.cpp b/tests/ktxdiff/ktxdiff_main.cpp index 30214767c3..6a8a988033 100644 --- a/tests/ktxdiff/ktxdiff_main.cpp +++ b/tests/ktxdiff/ktxdiff_main.cpp @@ -8,7 +8,7 @@ #include "vkformat_enum.h" #include "platform_utils.h" -#include "astc-encoder/Source/astcenc.h" +#include #include #include diff --git a/tests/transcodetests/CMakeLists.txt b/tests/transcodetests/CMakeLists.txt index 95d8e75334..2d98739422 100644 --- a/tests/transcodetests/CMakeLists.txt +++ b/tests/transcodetests/CMakeLists.txt @@ -19,7 +19,7 @@ target_link_libraries( gtest ktx obj_basisu_cbind - basisu_encoder + basisu::basisu_encoder ${CMAKE_THREAD_LIBS_INIT} ) diff --git a/tools/imageio/CMakeLists.txt b/tools/imageio/CMakeLists.txt index 32c8231b70..d4572cb27f 100644 --- a/tools/imageio/CMakeLists.txt +++ b/tools/imageio/CMakeLists.txt @@ -1,6 +1,9 @@ # Copyright 2022 The Khronos Group Inc. # SPDX-License-Identifier: Apache-2.0 +include("../../external/lodepng.cmake") +include("../../external/basisu.cmake") + set( PLUGIN_SOURCES exr.imageio/exrinput.cc jpg.imageio/jpginput.cc @@ -18,22 +21,15 @@ set( EXTERNAL set( EXR_PLUGIN_EXTERNAL_SOURCES ) set( EXR_PLUGIN_EXTERNAL_HEADERS - ${EXTERNAL}/astc-encoder/Source/ThirdParty/tinyexr.h + ${CMAKE_CURRENT_SOURCE_DIR}/external/tinyexr.h ) set( JPG_PLUGIN_EXTERNAL_SOURCES - ${EXTERNAL}/basis_universal/encoder/jpgd.cpp + ${CMAKE_CURRENT_SOURCE_DIR}/external/jpgd.cpp ) set( JPG_PLUGIN_EXTERNAL_HEADERS - ${EXTERNAL}/basis_universal/encoder/jpgd.h -) -set( PNG_PLUGIN_EXTERNAL_SOURCES - ${EXTERNAL}/lodepng/lodepng.cpp -) -set( PNG_PLUGIN_EXTERNAL_HEADERS - ${EXTERNAL}/lodepng/lodepng.h + ${CMAKE_CURRENT_SOURCE_DIR}/external/jpgd.h ) - add_library( imageio STATIC formatdesc.h image.hpp @@ -50,8 +46,6 @@ add_library( imageio STATIC ${EXR_PLUGIN_EXTERNAL_SOURCES} ${JPG_PLUGIN_EXTERNAL_HEADERS} ${JPG_PLUGIN_EXTERNAL_SOURCES} - ${PNG_PLUGIN_EXTERNAL_HEADERS} - ${PNG_PLUGIN_EXTERNAL_SOURCES} ) source_group(TREE "${CMAKE_CURRENT_LIST_DIR}" PREFIX "Source Files" FILES ${PLUGIN_SOURCES}) @@ -89,9 +83,8 @@ target_include_directories( PUBLIC ${PROJECT_SOURCE_DIR}/other_include PRIVATE - "${PROJECT_SOURCE_DIR}/external/astc-encoder/Source/ThirdParty" "${PROJECT_SOURCE_DIR}/external/dfdutils" - "${PROJECT_SOURCE_DIR}/external/lodepng" + "${CMAKE_CURRENT_SOURCE_DIR}/external" ) target_compile_definitions( @@ -102,7 +95,7 @@ PRIVATE target_compile_features(imageio PUBLIC cxx_std_17) -target_link_libraries(imageio fmt::fmt basisu_encoder) +target_link_libraries(imageio fmt::fmt basisu::basisu_encoder lodepng::lodepng) set_target_properties(imageio PROPERTIES CXX_VISIBILITY_PRESET ${STATIC_APP_LIB_SYMBOL_VISIBILITY} diff --git a/tools/imageio/external/jpgd.cpp b/tools/imageio/external/jpgd.cpp new file mode 100644 index 0000000000..57c7ec7b68 --- /dev/null +++ b/tools/imageio/external/jpgd.cpp @@ -0,0 +1,3230 @@ +// jpgd.cpp - C++ class for JPEG decompression. Written by Richard Geldreich between 1994-2020. +// Supports progressive and baseline sequential JPEG image files, and the most common chroma subsampling factors: Y, H1V1, H2V1, H1V2, and H2V2. +// Supports box and linear chroma upsampling. +// +// Released under two licenses. You are free to choose which license you want: +// License 1: +// Public Domain +// +// License 2: +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. +// +// Alex Evans: Linear memory allocator (taken from jpge.h). +// v1.04, May. 19, 2012: Code tweaks to fix VS2008 static code analysis warnings +// v2.00, March 20, 2020: Fuzzed with zzuf and afl. Fixed several issues, converted most assert()'s to run-time checks. Added chroma upsampling. Removed freq. domain upsampling. gcc/clang warnings. +// + +#include "jpgd.h" +#include +#include +#include + +#ifdef _MSC_VER +#pragma warning (disable : 4611) // warning C4611: interaction between '_setjmp' and C++ object destruction is non-portable +#endif + +#define JPGD_TRUE (1) +#define JPGD_FALSE (0) + +#define JPGD_MAX(a,b) (((a)>(b)) ? (a) : (b)) +#define JPGD_MIN(a,b) (((a)<(b)) ? (a) : (b)) + +namespace jpgd { + + static inline void* jpgd_malloc(size_t nSize) { return malloc(nSize); } + static inline void jpgd_free(void* p) { free(p); } + + // DCT coefficients are stored in this sequence. + static int g_ZAG[64] = { 0,1,8,16,9,2,3,10,17,24,32,25,18,11,4,5,12,19,26,33,40,48,41,34,27,20,13,6,7,14,21,28,35,42,49,56,57,50,43,36,29,22,15,23,30,37,44,51,58,59,52,45,38,31,39,46,53,60,61,54,47,55,62,63 }; + + enum JPEG_MARKER + { + M_SOF0 = 0xC0, M_SOF1 = 0xC1, M_SOF2 = 0xC2, M_SOF3 = 0xC3, M_SOF5 = 0xC5, M_SOF6 = 0xC6, M_SOF7 = 0xC7, M_JPG = 0xC8, + M_SOF9 = 0xC9, M_SOF10 = 0xCA, M_SOF11 = 0xCB, M_SOF13 = 0xCD, M_SOF14 = 0xCE, M_SOF15 = 0xCF, M_DHT = 0xC4, M_DAC = 0xCC, + M_RST0 = 0xD0, M_RST1 = 0xD1, M_RST2 = 0xD2, M_RST3 = 0xD3, M_RST4 = 0xD4, M_RST5 = 0xD5, M_RST6 = 0xD6, M_RST7 = 0xD7, + M_SOI = 0xD8, M_EOI = 0xD9, M_SOS = 0xDA, M_DQT = 0xDB, M_DNL = 0xDC, M_DRI = 0xDD, M_DHP = 0xDE, M_EXP = 0xDF, + M_APP0 = 0xE0, M_APP15 = 0xEF, M_JPG0 = 0xF0, M_JPG13 = 0xFD, M_COM = 0xFE, M_TEM = 0x01, M_ERROR = 0x100, RST0 = 0xD0 + }; + + enum JPEG_SUBSAMPLING { JPGD_GRAYSCALE = 0, JPGD_YH1V1, JPGD_YH2V1, JPGD_YH1V2, JPGD_YH2V2 }; + +#define CONST_BITS 13 +#define PASS1_BITS 2 +#define SCALEDONE ((int32)1) + +#define FIX_0_298631336 ((int32)2446) /* FIX(0.298631336) */ +#define FIX_0_390180644 ((int32)3196) /* FIX(0.390180644) */ +#define FIX_0_541196100 ((int32)4433) /* FIX(0.541196100) */ +#define FIX_0_765366865 ((int32)6270) /* FIX(0.765366865) */ +#define FIX_0_899976223 ((int32)7373) /* FIX(0.899976223) */ +#define FIX_1_175875602 ((int32)9633) /* FIX(1.175875602) */ +#define FIX_1_501321110 ((int32)12299) /* FIX(1.501321110) */ +#define FIX_1_847759065 ((int32)15137) /* FIX(1.847759065) */ +#define FIX_1_961570560 ((int32)16069) /* FIX(1.961570560) */ +#define FIX_2_053119869 ((int32)16819) /* FIX(2.053119869) */ +#define FIX_2_562915447 ((int32)20995) /* FIX(2.562915447) */ +#define FIX_3_072711026 ((int32)25172) /* FIX(3.072711026) */ + +#define DESCALE(x,n) (((x) + (SCALEDONE << ((n)-1))) >> (n)) +#define DESCALE_ZEROSHIFT(x,n) (((x) + (128 << (n)) + (SCALEDONE << ((n)-1))) >> (n)) + +#define MULTIPLY(var, cnst) ((var) * (cnst)) + +#define CLAMP(i) ((static_cast(i) > 255) ? (((~i) >> 31) & 0xFF) : (i)) + + static inline int left_shifti(int val, uint32_t bits) + { + return static_cast(static_cast(val) << bits); + } + + // Compiler creates a fast path 1D IDCT for X non-zero columns + template + struct Row + { + static void idct(int* pTemp, const jpgd_block_t* pSrc) + { + // ACCESS_COL() will be optimized at compile time to either an array access, or 0. Good compilers will then optimize out muls against 0. +#define ACCESS_COL(x) (((x) < NONZERO_COLS) ? (int)pSrc[x] : 0) + + const int z2 = ACCESS_COL(2), z3 = ACCESS_COL(6); + + const int z1 = MULTIPLY(z2 + z3, FIX_0_541196100); + const int tmp2 = z1 + MULTIPLY(z3, -FIX_1_847759065); + const int tmp3 = z1 + MULTIPLY(z2, FIX_0_765366865); + + const int tmp0 = left_shifti(ACCESS_COL(0) + ACCESS_COL(4), CONST_BITS); + const int tmp1 = left_shifti(ACCESS_COL(0) - ACCESS_COL(4), CONST_BITS); + + const int tmp10 = tmp0 + tmp3, tmp13 = tmp0 - tmp3, tmp11 = tmp1 + tmp2, tmp12 = tmp1 - tmp2; + + const int atmp0 = ACCESS_COL(7), atmp1 = ACCESS_COL(5), atmp2 = ACCESS_COL(3), atmp3 = ACCESS_COL(1); + + const int bz1 = atmp0 + atmp3, bz2 = atmp1 + atmp2, bz3 = atmp0 + atmp2, bz4 = atmp1 + atmp3; + const int bz5 = MULTIPLY(bz3 + bz4, FIX_1_175875602); + + const int az1 = MULTIPLY(bz1, -FIX_0_899976223); + const int az2 = MULTIPLY(bz2, -FIX_2_562915447); + const int az3 = MULTIPLY(bz3, -FIX_1_961570560) + bz5; + const int az4 = MULTIPLY(bz4, -FIX_0_390180644) + bz5; + + const int btmp0 = MULTIPLY(atmp0, FIX_0_298631336) + az1 + az3; + const int btmp1 = MULTIPLY(atmp1, FIX_2_053119869) + az2 + az4; + const int btmp2 = MULTIPLY(atmp2, FIX_3_072711026) + az2 + az3; + const int btmp3 = MULTIPLY(atmp3, FIX_1_501321110) + az1 + az4; + + pTemp[0] = DESCALE(tmp10 + btmp3, CONST_BITS - PASS1_BITS); + pTemp[7] = DESCALE(tmp10 - btmp3, CONST_BITS - PASS1_BITS); + pTemp[1] = DESCALE(tmp11 + btmp2, CONST_BITS - PASS1_BITS); + pTemp[6] = DESCALE(tmp11 - btmp2, CONST_BITS - PASS1_BITS); + pTemp[2] = DESCALE(tmp12 + btmp1, CONST_BITS - PASS1_BITS); + pTemp[5] = DESCALE(tmp12 - btmp1, CONST_BITS - PASS1_BITS); + pTemp[3] = DESCALE(tmp13 + btmp0, CONST_BITS - PASS1_BITS); + pTemp[4] = DESCALE(tmp13 - btmp0, CONST_BITS - PASS1_BITS); + } + }; + + template <> + struct Row<0> + { + static void idct(int* pTemp, const jpgd_block_t* pSrc) + { + (void)pTemp; + (void)pSrc; + } + }; + + template <> + struct Row<1> + { + static void idct(int* pTemp, const jpgd_block_t* pSrc) + { + const int dcval = left_shifti(pSrc[0], PASS1_BITS); + + pTemp[0] = dcval; + pTemp[1] = dcval; + pTemp[2] = dcval; + pTemp[3] = dcval; + pTemp[4] = dcval; + pTemp[5] = dcval; + pTemp[6] = dcval; + pTemp[7] = dcval; + } + }; + + // Compiler creates a fast path 1D IDCT for X non-zero rows + template + struct Col + { + static void idct(uint8* pDst_ptr, const int* pTemp) + { + // ACCESS_ROW() will be optimized at compile time to either an array access, or 0. +#define ACCESS_ROW(x) (((x) < NONZERO_ROWS) ? pTemp[x * 8] : 0) + + const int z2 = ACCESS_ROW(2); + const int z3 = ACCESS_ROW(6); + + const int z1 = MULTIPLY(z2 + z3, FIX_0_541196100); + const int tmp2 = z1 + MULTIPLY(z3, -FIX_1_847759065); + const int tmp3 = z1 + MULTIPLY(z2, FIX_0_765366865); + + const int tmp0 = left_shifti(ACCESS_ROW(0) + ACCESS_ROW(4), CONST_BITS); + const int tmp1 = left_shifti(ACCESS_ROW(0) - ACCESS_ROW(4), CONST_BITS); + + const int tmp10 = tmp0 + tmp3, tmp13 = tmp0 - tmp3, tmp11 = tmp1 + tmp2, tmp12 = tmp1 - tmp2; + + const int atmp0 = ACCESS_ROW(7), atmp1 = ACCESS_ROW(5), atmp2 = ACCESS_ROW(3), atmp3 = ACCESS_ROW(1); + + const int bz1 = atmp0 + atmp3, bz2 = atmp1 + atmp2, bz3 = atmp0 + atmp2, bz4 = atmp1 + atmp3; + const int bz5 = MULTIPLY(bz3 + bz4, FIX_1_175875602); + + const int az1 = MULTIPLY(bz1, -FIX_0_899976223); + const int az2 = MULTIPLY(bz2, -FIX_2_562915447); + const int az3 = MULTIPLY(bz3, -FIX_1_961570560) + bz5; + const int az4 = MULTIPLY(bz4, -FIX_0_390180644) + bz5; + + const int btmp0 = MULTIPLY(atmp0, FIX_0_298631336) + az1 + az3; + const int btmp1 = MULTIPLY(atmp1, FIX_2_053119869) + az2 + az4; + const int btmp2 = MULTIPLY(atmp2, FIX_3_072711026) + az2 + az3; + const int btmp3 = MULTIPLY(atmp3, FIX_1_501321110) + az1 + az4; + + int i = DESCALE_ZEROSHIFT(tmp10 + btmp3, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 0] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp10 - btmp3, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 7] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp11 + btmp2, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 1] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp11 - btmp2, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 6] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp12 + btmp1, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 2] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp12 - btmp1, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 5] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp13 + btmp0, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 3] = (uint8)CLAMP(i); + + i = DESCALE_ZEROSHIFT(tmp13 - btmp0, CONST_BITS + PASS1_BITS + 3); + pDst_ptr[8 * 4] = (uint8)CLAMP(i); + } + }; + + template <> + struct Col<1> + { + static void idct(uint8* pDst_ptr, const int* pTemp) + { + int dcval = DESCALE_ZEROSHIFT(pTemp[0], PASS1_BITS + 3); + const uint8 dcval_clamped = (uint8)CLAMP(dcval); + pDst_ptr[0 * 8] = dcval_clamped; + pDst_ptr[1 * 8] = dcval_clamped; + pDst_ptr[2 * 8] = dcval_clamped; + pDst_ptr[3 * 8] = dcval_clamped; + pDst_ptr[4 * 8] = dcval_clamped; + pDst_ptr[5 * 8] = dcval_clamped; + pDst_ptr[6 * 8] = dcval_clamped; + pDst_ptr[7 * 8] = dcval_clamped; + } + }; + + static const uint8 s_idct_row_table[] = + { + 1,0,0,0,0,0,0,0, 2,0,0,0,0,0,0,0, 2,1,0,0,0,0,0,0, 2,1,1,0,0,0,0,0, 2,2,1,0,0,0,0,0, 3,2,1,0,0,0,0,0, 4,2,1,0,0,0,0,0, 4,3,1,0,0,0,0,0, + 4,3,2,0,0,0,0,0, 4,3,2,1,0,0,0,0, 4,3,2,1,1,0,0,0, 4,3,2,2,1,0,0,0, 4,3,3,2,1,0,0,0, 4,4,3,2,1,0,0,0, 5,4,3,2,1,0,0,0, 6,4,3,2,1,0,0,0, + 6,5,3,2,1,0,0,0, 6,5,4,2,1,0,0,0, 6,5,4,3,1,0,0,0, 6,5,4,3,2,0,0,0, 6,5,4,3,2,1,0,0, 6,5,4,3,2,1,1,0, 6,5,4,3,2,2,1,0, 6,5,4,3,3,2,1,0, + 6,5,4,4,3,2,1,0, 6,5,5,4,3,2,1,0, 6,6,5,4,3,2,1,0, 7,6,5,4,3,2,1,0, 8,6,5,4,3,2,1,0, 8,7,5,4,3,2,1,0, 8,7,6,4,3,2,1,0, 8,7,6,5,3,2,1,0, + 8,7,6,5,4,2,1,0, 8,7,6,5,4,3,1,0, 8,7,6,5,4,3,2,0, 8,7,6,5,4,3,2,1, 8,7,6,5,4,3,2,2, 8,7,6,5,4,3,3,2, 8,7,6,5,4,4,3,2, 8,7,6,5,5,4,3,2, + 8,7,6,6,5,4,3,2, 8,7,7,6,5,4,3,2, 8,8,7,6,5,4,3,2, 8,8,8,6,5,4,3,2, 8,8,8,7,5,4,3,2, 8,8,8,7,6,4,3,2, 8,8,8,7,6,5,3,2, 8,8,8,7,6,5,4,2, + 8,8,8,7,6,5,4,3, 8,8,8,7,6,5,4,4, 8,8,8,7,6,5,5,4, 8,8,8,7,6,6,5,4, 8,8,8,7,7,6,5,4, 8,8,8,8,7,6,5,4, 8,8,8,8,8,6,5,4, 8,8,8,8,8,7,5,4, + 8,8,8,8,8,7,6,4, 8,8,8,8,8,7,6,5, 8,8,8,8,8,7,6,6, 8,8,8,8,8,7,7,6, 8,8,8,8,8,8,7,6, 8,8,8,8,8,8,8,6, 8,8,8,8,8,8,8,7, 8,8,8,8,8,8,8,8, + }; + + static const uint8 s_idct_col_table[] = + { + 1, 1, 2, 3, 3, 3, 3, 3, 3, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8 + }; + + // Scalar "fast pathing" IDCT. + static void idct(const jpgd_block_t* pSrc_ptr, uint8* pDst_ptr, int block_max_zag) + { + assert(block_max_zag >= 1); + assert(block_max_zag <= 64); + + if (block_max_zag <= 1) + { + int k = ((pSrc_ptr[0] + 4) >> 3) + 128; + k = CLAMP(k); + k = k | (k << 8); + k = k | (k << 16); + + for (int i = 8; i > 0; i--) + { + *(int*)&pDst_ptr[0] = k; + *(int*)&pDst_ptr[4] = k; + pDst_ptr += 8; + } + return; + } + + int temp[64]; + + const jpgd_block_t* pSrc = pSrc_ptr; + int* pTemp = temp; + + const uint8* pRow_tab = &s_idct_row_table[(block_max_zag - 1) * 8]; + int i; + for (i = 8; i > 0; i--, pRow_tab++) + { + switch (*pRow_tab) + { + case 0: Row<0>::idct(pTemp, pSrc); break; + case 1: Row<1>::idct(pTemp, pSrc); break; + case 2: Row<2>::idct(pTemp, pSrc); break; + case 3: Row<3>::idct(pTemp, pSrc); break; + case 4: Row<4>::idct(pTemp, pSrc); break; + case 5: Row<5>::idct(pTemp, pSrc); break; + case 6: Row<6>::idct(pTemp, pSrc); break; + case 7: Row<7>::idct(pTemp, pSrc); break; + case 8: Row<8>::idct(pTemp, pSrc); break; + } + + pSrc += 8; + pTemp += 8; + } + + pTemp = temp; + + const int nonzero_rows = s_idct_col_table[block_max_zag - 1]; + for (i = 8; i > 0; i--) + { + switch (nonzero_rows) + { + case 1: Col<1>::idct(pDst_ptr, pTemp); break; + case 2: Col<2>::idct(pDst_ptr, pTemp); break; + case 3: Col<3>::idct(pDst_ptr, pTemp); break; + case 4: Col<4>::idct(pDst_ptr, pTemp); break; + case 5: Col<5>::idct(pDst_ptr, pTemp); break; + case 6: Col<6>::idct(pDst_ptr, pTemp); break; + case 7: Col<7>::idct(pDst_ptr, pTemp); break; + case 8: Col<8>::idct(pDst_ptr, pTemp); break; + } + + pTemp++; + pDst_ptr++; + } + } + + // Retrieve one character from the input stream. + inline uint jpeg_decoder::get_char() + { + // Any bytes remaining in buffer? + if (!m_in_buf_left) + { + // Try to get more bytes. + prep_in_buffer(); + // Still nothing to get? + if (!m_in_buf_left) + { + // Pad the end of the stream with 0xFF 0xD9 (EOI marker) + int t = m_tem_flag; + m_tem_flag ^= 1; + if (t) + return 0xD9; + else + return 0xFF; + } + } + + uint c = *m_pIn_buf_ofs++; + m_in_buf_left--; + + return c; + } + + // Same as previous method, except can indicate if the character is a pad character or not. + inline uint jpeg_decoder::get_char(bool* pPadding_flag) + { + if (!m_in_buf_left) + { + prep_in_buffer(); + if (!m_in_buf_left) + { + *pPadding_flag = true; + int t = m_tem_flag; + m_tem_flag ^= 1; + if (t) + return 0xD9; + else + return 0xFF; + } + } + + *pPadding_flag = false; + + uint c = *m_pIn_buf_ofs++; + m_in_buf_left--; + + return c; + } + + // Inserts a previously retrieved character back into the input buffer. + inline void jpeg_decoder::stuff_char(uint8 q) + { + // This could write before the input buffer, but we've placed another array there. + *(--m_pIn_buf_ofs) = q; + m_in_buf_left++; + } + + // Retrieves one character from the input stream, but does not read past markers. Will continue to return 0xFF when a marker is encountered. + inline uint8 jpeg_decoder::get_octet() + { + bool padding_flag; + int c = get_char(&padding_flag); + + if (c == 0xFF) + { + if (padding_flag) + return 0xFF; + + c = get_char(&padding_flag); + if (padding_flag) + { + stuff_char(0xFF); + return 0xFF; + } + + if (c == 0x00) + return 0xFF; + else + { + stuff_char(static_cast(c)); + stuff_char(0xFF); + return 0xFF; + } + } + + return static_cast(c); + } + + // Retrieves a variable number of bits from the input stream. Does not recognize markers. + inline uint jpeg_decoder::get_bits(int num_bits) + { + if (!num_bits) + return 0; + + uint i = m_bit_buf >> (32 - num_bits); + + if ((m_bits_left -= num_bits) <= 0) + { + m_bit_buf <<= (num_bits += m_bits_left); + + uint c1 = get_char(); + uint c2 = get_char(); + m_bit_buf = (m_bit_buf & 0xFFFF0000) | (c1 << 8) | c2; + + m_bit_buf <<= -m_bits_left; + + m_bits_left += 16; + + assert(m_bits_left >= 0); + } + else + m_bit_buf <<= num_bits; + + return i; + } + + // Retrieves a variable number of bits from the input stream. Markers will not be read into the input bit buffer. Instead, an infinite number of all 1's will be returned when a marker is encountered. + inline uint jpeg_decoder::get_bits_no_markers(int num_bits) + { + if (!num_bits) + return 0; + + assert(num_bits <= 16); + + uint i = m_bit_buf >> (32 - num_bits); + + if ((m_bits_left -= num_bits) <= 0) + { + m_bit_buf <<= (num_bits += m_bits_left); + + if ((m_in_buf_left < 2) || (m_pIn_buf_ofs[0] == 0xFF) || (m_pIn_buf_ofs[1] == 0xFF)) + { + uint c1 = get_octet(); + uint c2 = get_octet(); + m_bit_buf |= (c1 << 8) | c2; + } + else + { + m_bit_buf |= ((uint)m_pIn_buf_ofs[0] << 8) | m_pIn_buf_ofs[1]; + m_in_buf_left -= 2; + m_pIn_buf_ofs += 2; + } + + m_bit_buf <<= -m_bits_left; + + m_bits_left += 16; + + assert(m_bits_left >= 0); + } + else + m_bit_buf <<= num_bits; + + return i; + } + + // Decodes a Huffman encoded symbol. + inline int jpeg_decoder::huff_decode(huff_tables* pH) + { + if (!pH) + stop_decoding(JPGD_DECODE_ERROR); + + int symbol; + // Check first 8-bits: do we have a complete symbol? + if ((symbol = pH->look_up[m_bit_buf >> 24]) < 0) + { + // Decode more bits, use a tree traversal to find symbol. + int ofs = 23; + do + { + unsigned int idx = -(int)(symbol + ((m_bit_buf >> ofs) & 1)); + + // This should never happen, but to be safe I'm turning these asserts into a run-time check. + if ((idx >= JPGD_HUFF_TREE_MAX_LENGTH) || (ofs < 0)) + stop_decoding(JPGD_DECODE_ERROR); + + symbol = pH->tree[idx]; + ofs--; + } while (symbol < 0); + + get_bits_no_markers(8 + (23 - ofs)); + } + else + { + assert(symbol < JPGD_HUFF_CODE_SIZE_MAX_LENGTH); + get_bits_no_markers(pH->code_size[symbol]); + } + + return symbol; + } + + // Decodes a Huffman encoded symbol. + inline int jpeg_decoder::huff_decode(huff_tables* pH, int& extra_bits) + { + int symbol; + + if (!pH) + stop_decoding(JPGD_DECODE_ERROR); + + // Check first 8-bits: do we have a complete symbol? + if ((symbol = pH->look_up2[m_bit_buf >> 24]) < 0) + { + // Use a tree traversal to find symbol. + int ofs = 23; + do + { + unsigned int idx = -(int)(symbol + ((m_bit_buf >> ofs) & 1)); + + // This should never happen, but to be safe I'm turning these asserts into a run-time check. + if ((idx >= JPGD_HUFF_TREE_MAX_LENGTH) || (ofs < 0)) + stop_decoding(JPGD_DECODE_ERROR); + + symbol = pH->tree[idx]; + ofs--; + } while (symbol < 0); + + get_bits_no_markers(8 + (23 - ofs)); + + extra_bits = get_bits_no_markers(symbol & 0xF); + } + else + { + if (symbol & 0x8000) + { + //get_bits_no_markers((symbol >> 8) & 31); + assert(((symbol >> 8) & 31) <= 15); + get_bits_no_markers((symbol >> 8) & 15); + extra_bits = symbol >> 16; + } + else + { + int code_size = (symbol >> 8) & 31; + int num_extra_bits = symbol & 0xF; + int bits = code_size + num_extra_bits; + + if (bits <= 16) + extra_bits = get_bits_no_markers(bits) & ((1 << num_extra_bits) - 1); + else + { + get_bits_no_markers(code_size); + extra_bits = get_bits_no_markers(num_extra_bits); + } + } + + symbol &= 0xFF; + } + + return symbol; + } + + // Tables and macro used to fully decode the DPCM differences. + static const int s_extend_test[16] = { 0, 0x0001, 0x0002, 0x0004, 0x0008, 0x0010, 0x0020, 0x0040, 0x0080, 0x0100, 0x0200, 0x0400, 0x0800, 0x1000, 0x2000, 0x4000 }; + static const int s_extend_offset[16] = { 0, -1, -3, -7, -15, -31, -63, -127, -255, -511, -1023, -2047, -4095, -8191, -16383, -32767 }; + //static const int s_extend_mask[] = { 0, (1 << 0), (1 << 1), (1 << 2), (1 << 3), (1 << 4), (1 << 5), (1 << 6), (1 << 7), (1 << 8), (1 << 9), (1 << 10), (1 << 11), (1 << 12), (1 << 13), (1 << 14), (1 << 15), (1 << 16) }; + +#define JPGD_HUFF_EXTEND(x, s) (((x) < s_extend_test[s & 15]) ? ((x) + s_extend_offset[s & 15]) : (x)) + + // Unconditionally frees all allocated m_blocks. + void jpeg_decoder::free_all_blocks() + { + m_pStream = nullptr; + for (mem_block* b = m_pMem_blocks; b; ) + { + mem_block* n = b->m_pNext; + jpgd_free(b); + b = n; + } + m_pMem_blocks = nullptr; + } + + // This method handles all errors. It will never return. + // It could easily be changed to use C++ exceptions. + JPGD_NORETURN void jpeg_decoder::stop_decoding(jpgd_status status) + { + m_error_code = status; + free_all_blocks(); + longjmp(m_jmp_state, status); + } + + void* jpeg_decoder::alloc(size_t nSize, bool zero) + { + nSize = (JPGD_MAX(nSize, 1) + 3) & ~3; + char* rv = nullptr; + for (mem_block* b = m_pMem_blocks; b; b = b->m_pNext) + { + if ((b->m_used_count + nSize) <= b->m_size) + { + rv = b->m_data + b->m_used_count; + b->m_used_count += nSize; + break; + } + } + if (!rv) + { + int capacity = JPGD_MAX(32768 - 256, ((int)nSize + 2047) & ~2047); + mem_block* b = (mem_block*)jpgd_malloc(sizeof(mem_block) + capacity); + if (!b) + { + stop_decoding(JPGD_NOTENOUGHMEM); + } + + b->m_pNext = m_pMem_blocks; + m_pMem_blocks = b; + b->m_used_count = nSize; + b->m_size = capacity; + rv = b->m_data; + } + if (zero) memset(rv, 0, nSize); + return rv; + } + + void jpeg_decoder::word_clear(void* p, uint16 c, uint n) + { + uint8* pD = (uint8*)p; + const uint8 l = c & 0xFF, h = (c >> 8) & 0xFF; + while (n) + { + pD[0] = l; + pD[1] = h; + pD += 2; + n--; + } + } + + // Refill the input buffer. + // This method will sit in a loop until (A) the buffer is full or (B) + // the stream's read() method reports and end of file condition. + void jpeg_decoder::prep_in_buffer() + { + m_in_buf_left = 0; + m_pIn_buf_ofs = m_in_buf; + + if (m_eof_flag) + return; + + do + { + int bytes_read = m_pStream->read(m_in_buf + m_in_buf_left, JPGD_IN_BUF_SIZE - m_in_buf_left, &m_eof_flag); + if (bytes_read == -1) + stop_decoding(JPGD_STREAM_READ); + + m_in_buf_left += bytes_read; + } while ((m_in_buf_left < JPGD_IN_BUF_SIZE) && (!m_eof_flag)); + + m_total_bytes_read += m_in_buf_left; + + // Pad the end of the block with M_EOI (prevents the decompressor from going off the rails if the stream is invalid). + // (This dates way back to when this decompressor was written in C/asm, and the all-asm Huffman decoder did some fancy things to increase perf.) + word_clear(m_pIn_buf_ofs + m_in_buf_left, 0xD9FF, 64); + } + + // Read a Huffman code table. + void jpeg_decoder::read_dht_marker() + { + int i, index, count; + uint8 huff_num[17]; + uint8 huff_val[256]; + + uint num_left = get_bits(16); + + if (num_left < 2) + stop_decoding(JPGD_BAD_DHT_MARKER); + + num_left -= 2; + + while (num_left) + { + index = get_bits(8); + + huff_num[0] = 0; + + count = 0; + + for (i = 1; i <= 16; i++) + { + huff_num[i] = static_cast(get_bits(8)); + count += huff_num[i]; + } + + if (count > 255) + stop_decoding(JPGD_BAD_DHT_COUNTS); + + bool symbol_present[256]; + memset(symbol_present, 0, sizeof(symbol_present)); + + for (i = 0; i < count; i++) + { + const int s = get_bits(8); + + // Check for obviously bogus tables. + if (symbol_present[s]) + stop_decoding(JPGD_BAD_DHT_COUNTS); + + huff_val[i] = static_cast(s); + symbol_present[s] = true; + } + + i = 1 + 16 + count; + + if (num_left < (uint)i) + stop_decoding(JPGD_BAD_DHT_MARKER); + + num_left -= i; + + if ((index & 0x10) > 0x10) + stop_decoding(JPGD_BAD_DHT_INDEX); + + index = (index & 0x0F) + ((index & 0x10) >> 4) * (JPGD_MAX_HUFF_TABLES >> 1); + + if (index >= JPGD_MAX_HUFF_TABLES) + stop_decoding(JPGD_BAD_DHT_INDEX); + + if (!m_huff_num[index]) + m_huff_num[index] = (uint8*)alloc(17); + + if (!m_huff_val[index]) + m_huff_val[index] = (uint8*)alloc(256); + + m_huff_ac[index] = (index & 0x10) != 0; + memcpy(m_huff_num[index], huff_num, 17); + memcpy(m_huff_val[index], huff_val, 256); + } + } + + // Read a quantization table. + void jpeg_decoder::read_dqt_marker() + { + int n, i, prec; + uint num_left; + uint temp; + + num_left = get_bits(16); + + if (num_left < 2) + stop_decoding(JPGD_BAD_DQT_MARKER); + + num_left -= 2; + + while (num_left) + { + n = get_bits(8); + prec = n >> 4; + n &= 0x0F; + + if (n >= JPGD_MAX_QUANT_TABLES) + stop_decoding(JPGD_BAD_DQT_TABLE); + + if (!m_quant[n]) + m_quant[n] = (jpgd_quant_t*)alloc(64 * sizeof(jpgd_quant_t)); + + // read quantization entries, in zag order + for (i = 0; i < 64; i++) + { + temp = get_bits(8); + + if (prec) + temp = (temp << 8) + get_bits(8); + + m_quant[n][i] = static_cast(temp); + } + + i = 64 + 1; + + if (prec) + i += 64; + + if (num_left < (uint)i) + stop_decoding(JPGD_BAD_DQT_LENGTH); + + num_left -= i; + } + } + + // Read the start of frame (SOF) marker. + void jpeg_decoder::read_sof_marker() + { + int i; + uint num_left; + + num_left = get_bits(16); + + /* precision: sorry, only 8-bit precision is supported */ + if (get_bits(8) != 8) + stop_decoding(JPGD_BAD_PRECISION); + + m_image_y_size = get_bits(16); + + if ((m_image_y_size < 1) || (m_image_y_size > JPGD_MAX_HEIGHT)) + stop_decoding(JPGD_BAD_HEIGHT); + + m_image_x_size = get_bits(16); + + if ((m_image_x_size < 1) || (m_image_x_size > JPGD_MAX_WIDTH)) + stop_decoding(JPGD_BAD_WIDTH); + + m_comps_in_frame = get_bits(8); + + if (m_comps_in_frame > JPGD_MAX_COMPONENTS) + stop_decoding(JPGD_TOO_MANY_COMPONENTS); + + if (num_left != (uint)(m_comps_in_frame * 3 + 8)) + stop_decoding(JPGD_BAD_SOF_LENGTH); + + for (i = 0; i < m_comps_in_frame; i++) + { + m_comp_ident[i] = get_bits(8); + m_comp_h_samp[i] = get_bits(4); + m_comp_v_samp[i] = get_bits(4); + + if (!m_comp_h_samp[i] || !m_comp_v_samp[i] || (m_comp_h_samp[i] > 2) || (m_comp_v_samp[i] > 2)) + stop_decoding(JPGD_UNSUPPORTED_SAMP_FACTORS); + + m_comp_quant[i] = get_bits(8); + if (m_comp_quant[i] >= JPGD_MAX_QUANT_TABLES) + stop_decoding(JPGD_DECODE_ERROR); + } + } + + // Used to skip unrecognized markers. + void jpeg_decoder::skip_variable_marker() + { + uint num_left; + + num_left = get_bits(16); + + if (num_left < 2) + stop_decoding(JPGD_BAD_VARIABLE_MARKER); + + num_left -= 2; + + while (num_left) + { + get_bits(8); + num_left--; + } + } + + // Read a define restart interval (DRI) marker. + void jpeg_decoder::read_dri_marker() + { + if (get_bits(16) != 4) + stop_decoding(JPGD_BAD_DRI_LENGTH); + + m_restart_interval = get_bits(16); + } + + // Read a start of scan (SOS) marker. + void jpeg_decoder::read_sos_marker() + { + uint num_left; + int i, ci, n, c, cc; + + num_left = get_bits(16); + + n = get_bits(8); + + m_comps_in_scan = n; + + num_left -= 3; + + if ((num_left != (uint)(n * 2 + 3)) || (n < 1) || (n > JPGD_MAX_COMPS_IN_SCAN)) + stop_decoding(JPGD_BAD_SOS_LENGTH); + + for (i = 0; i < n; i++) + { + cc = get_bits(8); + c = get_bits(8); + num_left -= 2; + + for (ci = 0; ci < m_comps_in_frame; ci++) + if (cc == m_comp_ident[ci]) + break; + + if (ci >= m_comps_in_frame) + stop_decoding(JPGD_BAD_SOS_COMP_ID); + + if (ci >= JPGD_MAX_COMPONENTS) + stop_decoding(JPGD_DECODE_ERROR); + + m_comp_list[i] = ci; + + m_comp_dc_tab[ci] = (c >> 4) & 15; + m_comp_ac_tab[ci] = (c & 15) + (JPGD_MAX_HUFF_TABLES >> 1); + + if (m_comp_dc_tab[ci] >= JPGD_MAX_HUFF_TABLES) + stop_decoding(JPGD_DECODE_ERROR); + + if (m_comp_ac_tab[ci] >= JPGD_MAX_HUFF_TABLES) + stop_decoding(JPGD_DECODE_ERROR); + } + + m_spectral_start = get_bits(8); + m_spectral_end = get_bits(8); + m_successive_high = get_bits(4); + m_successive_low = get_bits(4); + + if (!m_progressive_flag) + { + m_spectral_start = 0; + m_spectral_end = 63; + } + + num_left -= 3; + + /* read past whatever is num_left */ + while (num_left) + { + get_bits(8); + num_left--; + } + } + + // Finds the next marker. + int jpeg_decoder::next_marker() + { + uint c;// , bytes; + + //bytes = 0; + + do + { + do + { + //bytes++; + c = get_bits(8); + } while (c != 0xFF); + + do + { + c = get_bits(8); + } while (c == 0xFF); + + } while (c == 0); + + // If bytes > 0 here, there where extra bytes before the marker (not good). + + return c; + } + + // Process markers. Returns when an SOFx, SOI, EOI, or SOS marker is + // encountered. + int jpeg_decoder::process_markers() + { + int c; + + for (; ; ) + { + c = next_marker(); + + switch (c) + { + case M_SOF0: + case M_SOF1: + case M_SOF2: + case M_SOF3: + case M_SOF5: + case M_SOF6: + case M_SOF7: + // case M_JPG: + case M_SOF9: + case M_SOF10: + case M_SOF11: + case M_SOF13: + case M_SOF14: + case M_SOF15: + case M_SOI: + case M_EOI: + case M_SOS: + { + return c; + } + case M_DHT: + { + read_dht_marker(); + break; + } + // No arithmitic support - dumb patents! + case M_DAC: + { + stop_decoding(JPGD_NO_ARITHMITIC_SUPPORT); + break; + } + case M_DQT: + { + read_dqt_marker(); + break; + } + case M_DRI: + { + read_dri_marker(); + break; + } + //case M_APP0: /* no need to read the JFIF marker */ + case M_JPG: + case M_RST0: /* no parameters */ + case M_RST1: + case M_RST2: + case M_RST3: + case M_RST4: + case M_RST5: + case M_RST6: + case M_RST7: + case M_TEM: + { + stop_decoding(JPGD_UNEXPECTED_MARKER); + break; + } + default: /* must be DNL, DHP, EXP, APPn, JPGn, COM, or RESn or APP0 */ + { + skip_variable_marker(); + break; + } + } + } + } + + // Finds the start of image (SOI) marker. + void jpeg_decoder::locate_soi_marker() + { + uint lastchar, thischar; + uint bytesleft; + + lastchar = get_bits(8); + + thischar = get_bits(8); + + /* ok if it's a normal JPEG file without a special header */ + + if ((lastchar == 0xFF) && (thischar == M_SOI)) + return; + + bytesleft = 4096; + + for (; ; ) + { + if (--bytesleft == 0) + stop_decoding(JPGD_NOT_JPEG); + + lastchar = thischar; + + thischar = get_bits(8); + + if (lastchar == 0xFF) + { + if (thischar == M_SOI) + break; + else if (thischar == M_EOI) // get_bits will keep returning M_EOI if we read past the end + stop_decoding(JPGD_NOT_JPEG); + } + } + + // Check the next character after marker: if it's not 0xFF, it can't be the start of the next marker, so the file is bad. + thischar = (m_bit_buf >> 24) & 0xFF; + + if (thischar != 0xFF) + stop_decoding(JPGD_NOT_JPEG); + } + + // Find a start of frame (SOF) marker. + void jpeg_decoder::locate_sof_marker() + { + locate_soi_marker(); + + int c = process_markers(); + + switch (c) + { + case M_SOF2: + { + m_progressive_flag = JPGD_TRUE; + read_sof_marker(); + break; + } + case M_SOF0: /* baseline DCT */ + case M_SOF1: /* extended sequential DCT */ + { + read_sof_marker(); + break; + } + case M_SOF9: /* Arithmitic coding */ + { + stop_decoding(JPGD_NO_ARITHMITIC_SUPPORT); + break; + } + default: + { + stop_decoding(JPGD_UNSUPPORTED_MARKER); + break; + } + } + } + + // Find a start of scan (SOS) marker. + int jpeg_decoder::locate_sos_marker() + { + int c; + + c = process_markers(); + + if (c == M_EOI) + return JPGD_FALSE; + else if (c != M_SOS) + stop_decoding(JPGD_UNEXPECTED_MARKER); + + read_sos_marker(); + + return JPGD_TRUE; + } + + // Reset everything to default/uninitialized state. + void jpeg_decoder::init(jpeg_decoder_stream* pStream, uint32_t flags) + { + m_flags = flags; + m_pMem_blocks = nullptr; + m_error_code = JPGD_SUCCESS; + m_ready_flag = false; + m_image_x_size = m_image_y_size = 0; + m_pStream = pStream; + m_progressive_flag = JPGD_FALSE; + + memset(m_huff_ac, 0, sizeof(m_huff_ac)); + memset(m_huff_num, 0, sizeof(m_huff_num)); + memset(m_huff_val, 0, sizeof(m_huff_val)); + memset(m_quant, 0, sizeof(m_quant)); + + m_scan_type = 0; + m_comps_in_frame = 0; + + memset(m_comp_h_samp, 0, sizeof(m_comp_h_samp)); + memset(m_comp_v_samp, 0, sizeof(m_comp_v_samp)); + memset(m_comp_quant, 0, sizeof(m_comp_quant)); + memset(m_comp_ident, 0, sizeof(m_comp_ident)); + memset(m_comp_h_blocks, 0, sizeof(m_comp_h_blocks)); + memset(m_comp_v_blocks, 0, sizeof(m_comp_v_blocks)); + + m_comps_in_scan = 0; + memset(m_comp_list, 0, sizeof(m_comp_list)); + memset(m_comp_dc_tab, 0, sizeof(m_comp_dc_tab)); + memset(m_comp_ac_tab, 0, sizeof(m_comp_ac_tab)); + + m_spectral_start = 0; + m_spectral_end = 0; + m_successive_low = 0; + m_successive_high = 0; + m_max_mcu_x_size = 0; + m_max_mcu_y_size = 0; + m_blocks_per_mcu = 0; + m_max_blocks_per_row = 0; + m_mcus_per_row = 0; + m_mcus_per_col = 0; + + memset(m_mcu_org, 0, sizeof(m_mcu_org)); + + m_total_lines_left = 0; + m_mcu_lines_left = 0; + m_num_buffered_scanlines = 0; + m_real_dest_bytes_per_scan_line = 0; + m_dest_bytes_per_scan_line = 0; + m_dest_bytes_per_pixel = 0; + + memset(m_pHuff_tabs, 0, sizeof(m_pHuff_tabs)); + + memset(m_dc_coeffs, 0, sizeof(m_dc_coeffs)); + memset(m_ac_coeffs, 0, sizeof(m_ac_coeffs)); + memset(m_block_y_mcu, 0, sizeof(m_block_y_mcu)); + + m_eob_run = 0; + + m_pIn_buf_ofs = m_in_buf; + m_in_buf_left = 0; + m_eof_flag = false; + m_tem_flag = 0; + + memset(m_in_buf_pad_start, 0, sizeof(m_in_buf_pad_start)); + memset(m_in_buf, 0, sizeof(m_in_buf)); + memset(m_in_buf_pad_end, 0, sizeof(m_in_buf_pad_end)); + + m_restart_interval = 0; + m_restarts_left = 0; + m_next_restart_num = 0; + + m_max_mcus_per_row = 0; + m_max_blocks_per_mcu = 0; + m_max_mcus_per_col = 0; + + memset(m_last_dc_val, 0, sizeof(m_last_dc_val)); + m_pMCU_coefficients = nullptr; + m_pSample_buf = nullptr; + m_pSample_buf_prev = nullptr; + m_sample_buf_prev_valid = false; + + m_total_bytes_read = 0; + + m_pScan_line_0 = nullptr; + m_pScan_line_1 = nullptr; + + // Ready the input buffer. + prep_in_buffer(); + + // Prime the bit buffer. + m_bits_left = 16; + m_bit_buf = 0; + + get_bits(16); + get_bits(16); + + for (int i = 0; i < JPGD_MAX_BLOCKS_PER_MCU; i++) + m_mcu_block_max_zag[i] = 64; + } + +#define SCALEBITS 16 +#define ONE_HALF ((int) 1 << (SCALEBITS-1)) +#define FIX(x) ((int) ((x) * (1L<> SCALEBITS; + m_cbb[i] = (FIX(1.77200f) * k + ONE_HALF) >> SCALEBITS; + m_crg[i] = (-FIX(0.71414f)) * k; + m_cbg[i] = (-FIX(0.34414f)) * k + ONE_HALF; + } + } + + // This method throws back into the stream any bytes that where read + // into the bit buffer during initial marker scanning. + void jpeg_decoder::fix_in_buffer() + { + // In case any 0xFF's where pulled into the buffer during marker scanning. + assert((m_bits_left & 7) == 0); + + if (m_bits_left == 16) + stuff_char((uint8)(m_bit_buf & 0xFF)); + + if (m_bits_left >= 8) + stuff_char((uint8)((m_bit_buf >> 8) & 0xFF)); + + stuff_char((uint8)((m_bit_buf >> 16) & 0xFF)); + stuff_char((uint8)((m_bit_buf >> 24) & 0xFF)); + + m_bits_left = 16; + get_bits_no_markers(16); + get_bits_no_markers(16); + } + + void jpeg_decoder::transform_mcu(int mcu_row) + { + jpgd_block_t* pSrc_ptr = m_pMCU_coefficients; + if (mcu_row * m_blocks_per_mcu >= m_max_blocks_per_row) + stop_decoding(JPGD_DECODE_ERROR); + + uint8* pDst_ptr = m_pSample_buf + mcu_row * m_blocks_per_mcu * 64; + + for (int mcu_block = 0; mcu_block < m_blocks_per_mcu; mcu_block++) + { + idct(pSrc_ptr, pDst_ptr, m_mcu_block_max_zag[mcu_block]); + pSrc_ptr += 64; + pDst_ptr += 64; + } + } + + // Loads and dequantizes the next row of (already decoded) coefficients. + // Progressive images only. + void jpeg_decoder::load_next_row() + { + int i; + jpgd_block_t* p; + jpgd_quant_t* q; + int mcu_row, mcu_block;// , row_block = 0; + int component_num, component_id; + int block_x_mcu[JPGD_MAX_COMPONENTS]; + + memset(block_x_mcu, 0, JPGD_MAX_COMPONENTS * sizeof(int)); + + for (mcu_row = 0; mcu_row < m_mcus_per_row; mcu_row++) + { + int block_x_mcu_ofs = 0, block_y_mcu_ofs = 0; + + for (mcu_block = 0; mcu_block < m_blocks_per_mcu; mcu_block++) + { + component_id = m_mcu_org[mcu_block]; + if (m_comp_quant[component_id] >= JPGD_MAX_QUANT_TABLES) + stop_decoding(JPGD_DECODE_ERROR); + + q = m_quant[m_comp_quant[component_id]]; + + p = m_pMCU_coefficients + 64 * mcu_block; + + jpgd_block_t* pAC = coeff_buf_getp(m_ac_coeffs[component_id], block_x_mcu[component_id] + block_x_mcu_ofs, m_block_y_mcu[component_id] + block_y_mcu_ofs); + jpgd_block_t* pDC = coeff_buf_getp(m_dc_coeffs[component_id], block_x_mcu[component_id] + block_x_mcu_ofs, m_block_y_mcu[component_id] + block_y_mcu_ofs); + p[0] = pDC[0]; + memcpy(&p[1], &pAC[1], 63 * sizeof(jpgd_block_t)); + + for (i = 63; i > 0; i--) + if (p[g_ZAG[i]]) + break; + + m_mcu_block_max_zag[mcu_block] = i + 1; + + for (; i >= 0; i--) + if (p[g_ZAG[i]]) + p[g_ZAG[i]] = static_cast(p[g_ZAG[i]] * q[i]); + + //row_block++; + + if (m_comps_in_scan == 1) + block_x_mcu[component_id]++; + else + { + if (++block_x_mcu_ofs == m_comp_h_samp[component_id]) + { + block_x_mcu_ofs = 0; + + if (++block_y_mcu_ofs == m_comp_v_samp[component_id]) + { + block_y_mcu_ofs = 0; + + block_x_mcu[component_id] += m_comp_h_samp[component_id]; + } + } + } + } + + transform_mcu(mcu_row); + } + + if (m_comps_in_scan == 1) + m_block_y_mcu[m_comp_list[0]]++; + else + { + for (component_num = 0; component_num < m_comps_in_scan; component_num++) + { + component_id = m_comp_list[component_num]; + + m_block_y_mcu[component_id] += m_comp_v_samp[component_id]; + } + } + } + + // Restart interval processing. + void jpeg_decoder::process_restart() + { + int i; + int c = 0; + + // Align to a byte boundry + // FIXME: Is this really necessary? get_bits_no_markers() never reads in markers! + //get_bits_no_markers(m_bits_left & 7); + + // Let's scan a little bit to find the marker, but not _too_ far. + // 1536 is a "fudge factor" that determines how much to scan. + for (i = 1536; i > 0; i--) + if (get_char() == 0xFF) + break; + + if (i == 0) + stop_decoding(JPGD_BAD_RESTART_MARKER); + + for (; i > 0; i--) + if ((c = get_char()) != 0xFF) + break; + + if (i == 0) + stop_decoding(JPGD_BAD_RESTART_MARKER); + + // Is it the expected marker? If not, something bad happened. + if (c != (m_next_restart_num + M_RST0)) + stop_decoding(JPGD_BAD_RESTART_MARKER); + + // Reset each component's DC prediction values. + memset(&m_last_dc_val, 0, m_comps_in_frame * sizeof(uint)); + + m_eob_run = 0; + + m_restarts_left = m_restart_interval; + + m_next_restart_num = (m_next_restart_num + 1) & 7; + + // Get the bit buffer going again... + + m_bits_left = 16; + get_bits_no_markers(16); + get_bits_no_markers(16); + } + + static inline int dequantize_ac(int c, int q) { c *= q; return c; } + + // Decodes and dequantizes the next row of coefficients. + void jpeg_decoder::decode_next_row() + { + //int row_block = 0; + + for (int mcu_row = 0; mcu_row < m_mcus_per_row; mcu_row++) + { + if ((m_restart_interval) && (m_restarts_left == 0)) + process_restart(); + + jpgd_block_t* p = m_pMCU_coefficients; + for (int mcu_block = 0; mcu_block < m_blocks_per_mcu; mcu_block++, p += 64) + { + int component_id = m_mcu_org[mcu_block]; + if (m_comp_quant[component_id] >= JPGD_MAX_QUANT_TABLES) + stop_decoding(JPGD_DECODE_ERROR); + + jpgd_quant_t* q = m_quant[m_comp_quant[component_id]]; + + int r, s; + s = huff_decode(m_pHuff_tabs[m_comp_dc_tab[component_id]], r); + if (s >= 16) + stop_decoding(JPGD_DECODE_ERROR); + + s = JPGD_HUFF_EXTEND(r, s); + + m_last_dc_val[component_id] = (s += m_last_dc_val[component_id]); + + p[0] = static_cast(s * q[0]); + + int prev_num_set = m_mcu_block_max_zag[mcu_block]; + + huff_tables* pH = m_pHuff_tabs[m_comp_ac_tab[component_id]]; + + int k; + for (k = 1; k < 64; k++) + { + int extra_bits; + s = huff_decode(pH, extra_bits); + + r = s >> 4; + s &= 15; + + if (s) + { + if (r) + { + if ((k + r) > 63) + stop_decoding(JPGD_DECODE_ERROR); + + if (k < prev_num_set) + { + int n = JPGD_MIN(r, prev_num_set - k); + int kt = k; + while (n--) + p[g_ZAG[kt++]] = 0; + } + + k += r; + } + + s = JPGD_HUFF_EXTEND(extra_bits, s); + + if (k >= 64) + stop_decoding(JPGD_DECODE_ERROR); + + p[g_ZAG[k]] = static_cast(dequantize_ac(s, q[k])); //s * q[k]; + } + else + { + if (r == 15) + { + if ((k + 16) > 64) + stop_decoding(JPGD_DECODE_ERROR); + + if (k < prev_num_set) + { + int n = JPGD_MIN(16, prev_num_set - k); + int kt = k; + while (n--) + { + if (kt > 63) + stop_decoding(JPGD_DECODE_ERROR); + p[g_ZAG[kt++]] = 0; + } + } + + k += 16 - 1; // - 1 because the loop counter is k + + if (p[g_ZAG[k & 63]] != 0) + stop_decoding(JPGD_DECODE_ERROR); + } + else + break; + } + } + + if (k < prev_num_set) + { + int kt = k; + while (kt < prev_num_set) + p[g_ZAG[kt++]] = 0; + } + + m_mcu_block_max_zag[mcu_block] = k; + + //row_block++; + } + + transform_mcu(mcu_row); + + m_restarts_left--; + } + } + + // YCbCr H1V1 (1x1:1:1, 3 m_blocks per MCU) to RGB + void jpeg_decoder::H1V1Convert() + { + int row = m_max_mcu_y_size - m_mcu_lines_left; + uint8* d = m_pScan_line_0; + uint8* s = m_pSample_buf + row * 8; + + for (int i = m_max_mcus_per_row; i > 0; i--) + { + for (int j = 0; j < 8; j++) + { + int y = s[j]; + int cb = s[64 + j]; + int cr = s[128 + j]; + + d[0] = clamp(y + m_crr[cr]); + d[1] = clamp(y + ((m_crg[cr] + m_cbg[cb]) >> 16)); + d[2] = clamp(y + m_cbb[cb]); + d[3] = 255; + + d += 4; + } + + s += 64 * 3; + } + } + + // YCbCr H2V1 (2x1:1:1, 4 m_blocks per MCU) to RGB + void jpeg_decoder::H2V1Convert() + { + int row = m_max_mcu_y_size - m_mcu_lines_left; + uint8* d0 = m_pScan_line_0; + uint8* y = m_pSample_buf + row * 8; + uint8* c = m_pSample_buf + 2 * 64 + row * 8; + + for (int i = m_max_mcus_per_row; i > 0; i--) + { + for (int l = 0; l < 2; l++) + { + for (int j = 0; j < 4; j++) + { + int cb = c[0]; + int cr = c[64]; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + int yy = y[j << 1]; + d0[0] = clamp(yy + rc); + d0[1] = clamp(yy + gc); + d0[2] = clamp(yy + bc); + d0[3] = 255; + + yy = y[(j << 1) + 1]; + d0[4] = clamp(yy + rc); + d0[5] = clamp(yy + gc); + d0[6] = clamp(yy + bc); + d0[7] = 255; + + d0 += 8; + + c++; + } + y += 64; + } + + y += 64 * 4 - 64 * 2; + c += 64 * 4 - 8; + } + } + + // YCbCr H2V1 (2x1:1:1, 4 m_blocks per MCU) to RGB + void jpeg_decoder::H2V1ConvertFiltered() + { + const uint BLOCKS_PER_MCU = 4; + int row = m_max_mcu_y_size - m_mcu_lines_left; + uint8* d0 = m_pScan_line_0; + + const int half_image_x_size = (m_image_x_size >> 1) - 1; + const int row_x8 = row * 8; + + for (int x = 0; x < m_image_x_size; x++) + { + int y = m_pSample_buf[check_sample_buf_ofs((x >> 4) * BLOCKS_PER_MCU * 64 + ((x & 8) ? 64 : 0) + (x & 7) + row_x8)]; + + int c_x0 = (x - 1) >> 1; + int c_x1 = JPGD_MIN(c_x0 + 1, half_image_x_size); + c_x0 = JPGD_MAX(c_x0, 0); + + int a = (c_x0 >> 3) * BLOCKS_PER_MCU * 64 + (c_x0 & 7) + row_x8 + 128; + int cb0 = m_pSample_buf[check_sample_buf_ofs(a)]; + int cr0 = m_pSample_buf[check_sample_buf_ofs(a + 64)]; + + int b = (c_x1 >> 3) * BLOCKS_PER_MCU * 64 + (c_x1 & 7) + row_x8 + 128; + int cb1 = m_pSample_buf[check_sample_buf_ofs(b)]; + int cr1 = m_pSample_buf[check_sample_buf_ofs(b + 64)]; + + int w0 = (x & 1) ? 3 : 1; + int w1 = (x & 1) ? 1 : 3; + + int cb = (cb0 * w0 + cb1 * w1 + 2) >> 2; + int cr = (cr0 * w0 + cr1 * w1 + 2) >> 2; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d0[0] = clamp(y + rc); + d0[1] = clamp(y + gc); + d0[2] = clamp(y + bc); + d0[3] = 255; + + d0 += 4; + } + } + + // YCbCr H2V1 (1x2:1:1, 4 m_blocks per MCU) to RGB + void jpeg_decoder::H1V2Convert() + { + int row = m_max_mcu_y_size - m_mcu_lines_left; + uint8* d0 = m_pScan_line_0; + uint8* d1 = m_pScan_line_1; + uint8* y; + uint8* c; + + if (row < 8) + y = m_pSample_buf + row * 8; + else + y = m_pSample_buf + 64 * 1 + (row & 7) * 8; + + c = m_pSample_buf + 64 * 2 + (row >> 1) * 8; + + for (int i = m_max_mcus_per_row; i > 0; i--) + { + for (int j = 0; j < 8; j++) + { + int cb = c[0 + j]; + int cr = c[64 + j]; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + int yy = y[j]; + d0[0] = clamp(yy + rc); + d0[1] = clamp(yy + gc); + d0[2] = clamp(yy + bc); + d0[3] = 255; + + yy = y[8 + j]; + d1[0] = clamp(yy + rc); + d1[1] = clamp(yy + gc); + d1[2] = clamp(yy + bc); + d1[3] = 255; + + d0 += 4; + d1 += 4; + } + + y += 64 * 4; + c += 64 * 4; + } + } + + // YCbCr H2V1 (1x2:1:1, 4 m_blocks per MCU) to RGB + void jpeg_decoder::H1V2ConvertFiltered() + { + const uint BLOCKS_PER_MCU = 4; + int y = m_image_y_size - m_total_lines_left; + int row = y & 15; + + const int half_image_y_size = (m_image_y_size >> 1) - 1; + + uint8* d0 = m_pScan_line_0; + + const int w0 = (row & 1) ? 3 : 1; + const int w1 = (row & 1) ? 1 : 3; + + int c_y0 = (y - 1) >> 1; + int c_y1 = JPGD_MIN(c_y0 + 1, half_image_y_size); + + const uint8_t* p_YSamples = m_pSample_buf; + const uint8_t* p_C0Samples = m_pSample_buf; + if ((c_y0 >= 0) && (((row & 15) == 0) || ((row & 15) == 15)) && (m_total_lines_left > 1)) + { + assert(y > 0); + assert(m_sample_buf_prev_valid); + + if ((row & 15) == 15) + p_YSamples = m_pSample_buf_prev; + + p_C0Samples = m_pSample_buf_prev; + } + + const int y_sample_base_ofs = ((row & 8) ? 64 : 0) + (row & 7) * 8; + const int y0_base = (c_y0 & 7) * 8 + 128; + const int y1_base = (c_y1 & 7) * 8 + 128; + + for (int x = 0; x < m_image_x_size; x++) + { + const int base_ofs = (x >> 3) * BLOCKS_PER_MCU * 64 + (x & 7); + + int y_sample = p_YSamples[check_sample_buf_ofs(base_ofs + y_sample_base_ofs)]; + + int a = base_ofs + y0_base; + int cb0_sample = p_C0Samples[check_sample_buf_ofs(a)]; + int cr0_sample = p_C0Samples[check_sample_buf_ofs(a + 64)]; + + int b = base_ofs + y1_base; + int cb1_sample = m_pSample_buf[check_sample_buf_ofs(b)]; + int cr1_sample = m_pSample_buf[check_sample_buf_ofs(b + 64)]; + + int cb = (cb0_sample * w0 + cb1_sample * w1 + 2) >> 2; + int cr = (cr0_sample * w0 + cr1_sample * w1 + 2) >> 2; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d0[0] = clamp(y_sample + rc); + d0[1] = clamp(y_sample + gc); + d0[2] = clamp(y_sample + bc); + d0[3] = 255; + + d0 += 4; + } + } + + // YCbCr H2V2 (2x2:1:1, 6 m_blocks per MCU) to RGB + void jpeg_decoder::H2V2Convert() + { + int row = m_max_mcu_y_size - m_mcu_lines_left; + uint8* d0 = m_pScan_line_0; + uint8* d1 = m_pScan_line_1; + uint8* y; + uint8* c; + + if (row < 8) + y = m_pSample_buf + row * 8; + else + y = m_pSample_buf + 64 * 2 + (row & 7) * 8; + + c = m_pSample_buf + 64 * 4 + (row >> 1) * 8; + + for (int i = m_max_mcus_per_row; i > 0; i--) + { + for (int l = 0; l < 2; l++) + { + for (int j = 0; j < 8; j += 2) + { + int cb = c[0]; + int cr = c[64]; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + int yy = y[j]; + d0[0] = clamp(yy + rc); + d0[1] = clamp(yy + gc); + d0[2] = clamp(yy + bc); + d0[3] = 255; + + yy = y[j + 1]; + d0[4] = clamp(yy + rc); + d0[5] = clamp(yy + gc); + d0[6] = clamp(yy + bc); + d0[7] = 255; + + yy = y[j + 8]; + d1[0] = clamp(yy + rc); + d1[1] = clamp(yy + gc); + d1[2] = clamp(yy + bc); + d1[3] = 255; + + yy = y[j + 8 + 1]; + d1[4] = clamp(yy + rc); + d1[5] = clamp(yy + gc); + d1[6] = clamp(yy + bc); + d1[7] = 255; + + d0 += 8; + d1 += 8; + + c++; + } + y += 64; + } + + y += 64 * 6 - 64 * 2; + c += 64 * 6 - 8; + } + } + + uint32_t jpeg_decoder::H2V2ConvertFiltered() + { + const uint BLOCKS_PER_MCU = 6; + int y = m_image_y_size - m_total_lines_left; + int row = y & 15; + + const int half_image_y_size = (m_image_y_size >> 1) - 1; + + uint8* d0 = m_pScan_line_0; + + int c_y0 = (y - 1) >> 1; + int c_y1 = JPGD_MIN(c_y0 + 1, half_image_y_size); + + const uint8_t* p_YSamples = m_pSample_buf; + const uint8_t* p_C0Samples = m_pSample_buf; + if ((c_y0 >= 0) && (((row & 15) == 0) || ((row & 15) == 15)) && (m_total_lines_left > 1)) + { + assert(y > 0); + assert(m_sample_buf_prev_valid); + + if ((row & 15) == 15) + p_YSamples = m_pSample_buf_prev; + + p_C0Samples = m_pSample_buf_prev; + } + + const int y_sample_base_ofs = ((row & 8) ? 128 : 0) + (row & 7) * 8; + const int y0_base = (c_y0 & 7) * 8 + 256; + const int y1_base = (c_y1 & 7) * 8 + 256; + + const int half_image_x_size = (m_image_x_size >> 1) - 1; + + static const uint8_t s_muls[2][2][4] = + { + { { 1, 3, 3, 9 }, { 3, 9, 1, 3 }, }, + { { 3, 1, 9, 3 }, { 9, 3, 3, 1 } } + }; + + if (((row & 15) >= 1) && ((row & 15) <= 14)) + { + assert((row & 1) == 1); + assert(((y + 1 - 1) >> 1) == c_y0); + + assert(p_YSamples == m_pSample_buf); + assert(p_C0Samples == m_pSample_buf); + + uint8* d1 = m_pScan_line_1; + const int y_sample_base_ofs1 = (((row + 1) & 8) ? 128 : 0) + ((row + 1) & 7) * 8; + + for (int x = 0; x < m_image_x_size; x++) + { + int k = (x >> 4) * BLOCKS_PER_MCU * 64 + ((x & 8) ? 64 : 0) + (x & 7); + int y_sample0 = p_YSamples[check_sample_buf_ofs(k + y_sample_base_ofs)]; + int y_sample1 = p_YSamples[check_sample_buf_ofs(k + y_sample_base_ofs1)]; + + int c_x0 = (x - 1) >> 1; + int c_x1 = JPGD_MIN(c_x0 + 1, half_image_x_size); + c_x0 = JPGD_MAX(c_x0, 0); + + int a = (c_x0 >> 3) * BLOCKS_PER_MCU * 64 + (c_x0 & 7); + int cb00_sample = p_C0Samples[check_sample_buf_ofs(a + y0_base)]; + int cr00_sample = p_C0Samples[check_sample_buf_ofs(a + y0_base + 64)]; + + int cb01_sample = m_pSample_buf[check_sample_buf_ofs(a + y1_base)]; + int cr01_sample = m_pSample_buf[check_sample_buf_ofs(a + y1_base + 64)]; + + int b = (c_x1 >> 3) * BLOCKS_PER_MCU * 64 + (c_x1 & 7); + int cb10_sample = p_C0Samples[check_sample_buf_ofs(b + y0_base)]; + int cr10_sample = p_C0Samples[check_sample_buf_ofs(b + y0_base + 64)]; + + int cb11_sample = m_pSample_buf[check_sample_buf_ofs(b + y1_base)]; + int cr11_sample = m_pSample_buf[check_sample_buf_ofs(b + y1_base + 64)]; + + { + const uint8_t* pMuls = &s_muls[row & 1][x & 1][0]; + int cb = (cb00_sample * pMuls[0] + cb01_sample * pMuls[1] + cb10_sample * pMuls[2] + cb11_sample * pMuls[3] + 8) >> 4; + int cr = (cr00_sample * pMuls[0] + cr01_sample * pMuls[1] + cr10_sample * pMuls[2] + cr11_sample * pMuls[3] + 8) >> 4; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d0[0] = clamp(y_sample0 + rc); + d0[1] = clamp(y_sample0 + gc); + d0[2] = clamp(y_sample0 + bc); + d0[3] = 255; + + d0 += 4; + } + + { + const uint8_t* pMuls = &s_muls[(row + 1) & 1][x & 1][0]; + int cb = (cb00_sample * pMuls[0] + cb01_sample * pMuls[1] + cb10_sample * pMuls[2] + cb11_sample * pMuls[3] + 8) >> 4; + int cr = (cr00_sample * pMuls[0] + cr01_sample * pMuls[1] + cr10_sample * pMuls[2] + cr11_sample * pMuls[3] + 8) >> 4; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d1[0] = clamp(y_sample1 + rc); + d1[1] = clamp(y_sample1 + gc); + d1[2] = clamp(y_sample1 + bc); + d1[3] = 255; + + d1 += 4; + } + + if (((x & 1) == 1) && (x < m_image_x_size - 1)) + { + const int nx = x + 1; + assert(c_x0 == (nx - 1) >> 1); + + k = (nx >> 4) * BLOCKS_PER_MCU * 64 + ((nx & 8) ? 64 : 0) + (nx & 7); + y_sample0 = p_YSamples[check_sample_buf_ofs(k + y_sample_base_ofs)]; + y_sample1 = p_YSamples[check_sample_buf_ofs(k + y_sample_base_ofs1)]; + + { + const uint8_t* pMuls = &s_muls[row & 1][nx & 1][0]; + int cb = (cb00_sample * pMuls[0] + cb01_sample * pMuls[1] + cb10_sample * pMuls[2] + cb11_sample * pMuls[3] + 8) >> 4; + int cr = (cr00_sample * pMuls[0] + cr01_sample * pMuls[1] + cr10_sample * pMuls[2] + cr11_sample * pMuls[3] + 8) >> 4; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d0[0] = clamp(y_sample0 + rc); + d0[1] = clamp(y_sample0 + gc); + d0[2] = clamp(y_sample0 + bc); + d0[3] = 255; + + d0 += 4; + } + + { + const uint8_t* pMuls = &s_muls[(row + 1) & 1][nx & 1][0]; + int cb = (cb00_sample * pMuls[0] + cb01_sample * pMuls[1] + cb10_sample * pMuls[2] + cb11_sample * pMuls[3] + 8) >> 4; + int cr = (cr00_sample * pMuls[0] + cr01_sample * pMuls[1] + cr10_sample * pMuls[2] + cr11_sample * pMuls[3] + 8) >> 4; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d1[0] = clamp(y_sample1 + rc); + d1[1] = clamp(y_sample1 + gc); + d1[2] = clamp(y_sample1 + bc); + d1[3] = 255; + + d1 += 4; + } + + ++x; + } + } + + return 2; + } + else + { + for (int x = 0; x < m_image_x_size; x++) + { + int y_sample = p_YSamples[check_sample_buf_ofs((x >> 4) * BLOCKS_PER_MCU * 64 + ((x & 8) ? 64 : 0) + (x & 7) + y_sample_base_ofs)]; + + int c_x0 = (x - 1) >> 1; + int c_x1 = JPGD_MIN(c_x0 + 1, half_image_x_size); + c_x0 = JPGD_MAX(c_x0, 0); + + int a = (c_x0 >> 3) * BLOCKS_PER_MCU * 64 + (c_x0 & 7); + int cb00_sample = p_C0Samples[check_sample_buf_ofs(a + y0_base)]; + int cr00_sample = p_C0Samples[check_sample_buf_ofs(a + y0_base + 64)]; + + int cb01_sample = m_pSample_buf[check_sample_buf_ofs(a + y1_base)]; + int cr01_sample = m_pSample_buf[check_sample_buf_ofs(a + y1_base + 64)]; + + int b = (c_x1 >> 3) * BLOCKS_PER_MCU * 64 + (c_x1 & 7); + int cb10_sample = p_C0Samples[check_sample_buf_ofs(b + y0_base)]; + int cr10_sample = p_C0Samples[check_sample_buf_ofs(b + y0_base + 64)]; + + int cb11_sample = m_pSample_buf[check_sample_buf_ofs(b + y1_base)]; + int cr11_sample = m_pSample_buf[check_sample_buf_ofs(b + y1_base + 64)]; + + const uint8_t* pMuls = &s_muls[row & 1][x & 1][0]; + int cb = (cb00_sample * pMuls[0] + cb01_sample * pMuls[1] + cb10_sample * pMuls[2] + cb11_sample * pMuls[3] + 8) >> 4; + int cr = (cr00_sample * pMuls[0] + cr01_sample * pMuls[1] + cr10_sample * pMuls[2] + cr11_sample * pMuls[3] + 8) >> 4; + + int rc = m_crr[cr]; + int gc = ((m_crg[cr] + m_cbg[cb]) >> 16); + int bc = m_cbb[cb]; + + d0[0] = clamp(y_sample + rc); + d0[1] = clamp(y_sample + gc); + d0[2] = clamp(y_sample + bc); + d0[3] = 255; + + d0 += 4; + } + + return 1; + } + } + + // Y (1 block per MCU) to 8-bit grayscale + void jpeg_decoder::gray_convert() + { + int row = m_max_mcu_y_size - m_mcu_lines_left; + uint8* d = m_pScan_line_0; + uint8* s = m_pSample_buf + row * 8; + + for (int i = m_max_mcus_per_row; i > 0; i--) + { + *(uint*)d = *(uint*)s; + *(uint*)(&d[4]) = *(uint*)(&s[4]); + + s += 64; + d += 8; + } + } + + // Find end of image (EOI) marker, so we can return to the user the exact size of the input stream. + void jpeg_decoder::find_eoi() + { + if (!m_progressive_flag) + { + // Attempt to read the EOI marker. + //get_bits_no_markers(m_bits_left & 7); + + // Prime the bit buffer + m_bits_left = 16; + get_bits(16); + get_bits(16); + + // The next marker _should_ be EOI + process_markers(); + } + + m_total_bytes_read -= m_in_buf_left; + } + + int jpeg_decoder::decode_next_mcu_row() + { + if (setjmp(m_jmp_state)) + return JPGD_FAILED; + + const bool chroma_y_filtering = (m_flags & cFlagLinearChromaFiltering) && ((m_scan_type == JPGD_YH2V2) || (m_scan_type == JPGD_YH1V2)) && (m_image_x_size >= 2) && (m_image_y_size >= 2); + if (chroma_y_filtering) + { + std::swap(m_pSample_buf, m_pSample_buf_prev); + + m_sample_buf_prev_valid = true; + } + + if (m_progressive_flag) + load_next_row(); + else + decode_next_row(); + + // Find the EOI marker if that was the last row. + if (m_total_lines_left <= m_max_mcu_y_size) + find_eoi(); + + m_mcu_lines_left = m_max_mcu_y_size; + return 0; + } + + int jpeg_decoder::decode(const void** pScan_line, uint* pScan_line_len) + { + if ((m_error_code) || (!m_ready_flag)) + return JPGD_FAILED; + + if (m_total_lines_left == 0) + return JPGD_DONE; + + const bool chroma_y_filtering = (m_flags & cFlagLinearChromaFiltering) && ((m_scan_type == JPGD_YH2V2) || (m_scan_type == JPGD_YH1V2)) && (m_image_x_size >= 2) && (m_image_y_size >= 2); + + bool get_another_mcu_row = false; + bool got_mcu_early = false; + if (chroma_y_filtering) + { + if (m_total_lines_left == m_image_y_size) + get_another_mcu_row = true; + else if ((m_mcu_lines_left == 1) && (m_total_lines_left > 1)) + { + get_another_mcu_row = true; + got_mcu_early = true; + } + } + else + { + get_another_mcu_row = (m_mcu_lines_left == 0); + } + + if (get_another_mcu_row) + { + int status = decode_next_mcu_row(); + if (status != 0) + return status; + } + + switch (m_scan_type) + { + case JPGD_YH2V2: + { + if ((m_flags & cFlagLinearChromaFiltering) && (m_image_x_size >= 2) && (m_image_y_size >= 2)) + { + if (m_num_buffered_scanlines == 1) + { + *pScan_line = m_pScan_line_1; + } + else if (m_num_buffered_scanlines == 0) + { + m_num_buffered_scanlines = H2V2ConvertFiltered(); + *pScan_line = m_pScan_line_0; + } + + m_num_buffered_scanlines--; + } + else + { + if ((m_mcu_lines_left & 1) == 0) + { + H2V2Convert(); + *pScan_line = m_pScan_line_0; + } + else + *pScan_line = m_pScan_line_1; + } + + break; + } + case JPGD_YH2V1: + { + if ((m_flags & cFlagLinearChromaFiltering) && (m_image_x_size >= 2) && (m_image_y_size >= 2)) + H2V1ConvertFiltered(); + else + H2V1Convert(); + *pScan_line = m_pScan_line_0; + break; + } + case JPGD_YH1V2: + { + if (chroma_y_filtering) + { + H1V2ConvertFiltered(); + *pScan_line = m_pScan_line_0; + } + else + { + if ((m_mcu_lines_left & 1) == 0) + { + H1V2Convert(); + *pScan_line = m_pScan_line_0; + } + else + *pScan_line = m_pScan_line_1; + } + + break; + } + case JPGD_YH1V1: + { + H1V1Convert(); + *pScan_line = m_pScan_line_0; + break; + } + case JPGD_GRAYSCALE: + { + gray_convert(); + *pScan_line = m_pScan_line_0; + + break; + } + } + + *pScan_line_len = m_real_dest_bytes_per_scan_line; + + if (!got_mcu_early) + { + m_mcu_lines_left--; + } + + m_total_lines_left--; + + return JPGD_SUCCESS; + } + + // Creates the tables needed for efficient Huffman decoding. + void jpeg_decoder::make_huff_table(int index, huff_tables* pH) + { + int p, i, l, si; + uint8 huffsize[258]; + uint huffcode[258]; + uint code; + uint subtree; + int code_size; + int lastp; + int nextfreeentry; + int currententry; + + pH->ac_table = m_huff_ac[index] != 0; + + p = 0; + + for (l = 1; l <= 16; l++) + { + for (i = 1; i <= m_huff_num[index][l]; i++) + { + if (p >= 257) + stop_decoding(JPGD_DECODE_ERROR); + huffsize[p++] = static_cast(l); + } + } + + assert(p < 258); + huffsize[p] = 0; + + lastp = p; + + code = 0; + si = huffsize[0]; + p = 0; + + while (huffsize[p]) + { + while (huffsize[p] == si) + { + if (p >= 257) + stop_decoding(JPGD_DECODE_ERROR); + huffcode[p++] = code; + code++; + } + + code <<= 1; + si++; + } + + memset(pH->look_up, 0, sizeof(pH->look_up)); + memset(pH->look_up2, 0, sizeof(pH->look_up2)); + memset(pH->tree, 0, sizeof(pH->tree)); + memset(pH->code_size, 0, sizeof(pH->code_size)); + + nextfreeentry = -1; + + p = 0; + + while (p < lastp) + { + i = m_huff_val[index][p]; + + code = huffcode[p]; + code_size = huffsize[p]; + + assert(i < JPGD_HUFF_CODE_SIZE_MAX_LENGTH); + pH->code_size[i] = static_cast(code_size); + + if (code_size <= 8) + { + code <<= (8 - code_size); + + for (l = 1 << (8 - code_size); l > 0; l--) + { + if (code >= 256) + stop_decoding(JPGD_DECODE_ERROR); + + pH->look_up[code] = i; + + bool has_extrabits = false; + int extra_bits = 0; + int num_extra_bits = i & 15; + + int bits_to_fetch = code_size; + if (num_extra_bits) + { + int total_codesize = code_size + num_extra_bits; + if (total_codesize <= 8) + { + has_extrabits = true; + extra_bits = ((1 << num_extra_bits) - 1) & (code >> (8 - total_codesize)); + + if (extra_bits > 0x7FFF) + stop_decoding(JPGD_DECODE_ERROR); + + bits_to_fetch += num_extra_bits; + } + } + + if (!has_extrabits) + pH->look_up2[code] = i | (bits_to_fetch << 8); + else + pH->look_up2[code] = i | 0x8000 | (extra_bits << 16) | (bits_to_fetch << 8); + + code++; + } + } + else + { + subtree = (code >> (code_size - 8)) & 0xFF; + + currententry = pH->look_up[subtree]; + + if (currententry == 0) + { + pH->look_up[subtree] = currententry = nextfreeentry; + pH->look_up2[subtree] = currententry = nextfreeentry; + + nextfreeentry -= 2; + } + + code <<= (16 - (code_size - 8)); + + for (l = code_size; l > 9; l--) + { + if ((code & 0x8000) == 0) + currententry--; + + unsigned int idx = -currententry - 1; + + if (idx >= JPGD_HUFF_TREE_MAX_LENGTH) + stop_decoding(JPGD_DECODE_ERROR); + + if (pH->tree[idx] == 0) + { + pH->tree[idx] = nextfreeentry; + + currententry = nextfreeentry; + + nextfreeentry -= 2; + } + else + { + currententry = pH->tree[idx]; + } + + code <<= 1; + } + + if ((code & 0x8000) == 0) + currententry--; + + if ((-currententry - 1) >= JPGD_HUFF_TREE_MAX_LENGTH) + stop_decoding(JPGD_DECODE_ERROR); + + pH->tree[-currententry - 1] = i; + } + + p++; + } + } + + // Verifies the quantization tables needed for this scan are available. + void jpeg_decoder::check_quant_tables() + { + for (int i = 0; i < m_comps_in_scan; i++) + if (m_quant[m_comp_quant[m_comp_list[i]]] == nullptr) + stop_decoding(JPGD_UNDEFINED_QUANT_TABLE); + } + + // Verifies that all the Huffman tables needed for this scan are available. + void jpeg_decoder::check_huff_tables() + { + for (int i = 0; i < m_comps_in_scan; i++) + { + if ((m_spectral_start == 0) && (m_huff_num[m_comp_dc_tab[m_comp_list[i]]] == nullptr)) + stop_decoding(JPGD_UNDEFINED_HUFF_TABLE); + + if ((m_spectral_end > 0) && (m_huff_num[m_comp_ac_tab[m_comp_list[i]]] == nullptr)) + stop_decoding(JPGD_UNDEFINED_HUFF_TABLE); + } + + for (int i = 0; i < JPGD_MAX_HUFF_TABLES; i++) + if (m_huff_num[i]) + { + if (!m_pHuff_tabs[i]) + m_pHuff_tabs[i] = (huff_tables*)alloc(sizeof(huff_tables)); + + make_huff_table(i, m_pHuff_tabs[i]); + } + } + + // Determines the component order inside each MCU. + // Also calcs how many MCU's are on each row, etc. + bool jpeg_decoder::calc_mcu_block_order() + { + int component_num, component_id; + int max_h_samp = 0, max_v_samp = 0; + + for (component_id = 0; component_id < m_comps_in_frame; component_id++) + { + if (m_comp_h_samp[component_id] > max_h_samp) + max_h_samp = m_comp_h_samp[component_id]; + + if (m_comp_v_samp[component_id] > max_v_samp) + max_v_samp = m_comp_v_samp[component_id]; + } + + for (component_id = 0; component_id < m_comps_in_frame; component_id++) + { + m_comp_h_blocks[component_id] = ((((m_image_x_size * m_comp_h_samp[component_id]) + (max_h_samp - 1)) / max_h_samp) + 7) / 8; + m_comp_v_blocks[component_id] = ((((m_image_y_size * m_comp_v_samp[component_id]) + (max_v_samp - 1)) / max_v_samp) + 7) / 8; + } + + if (m_comps_in_scan == 1) + { + m_mcus_per_row = m_comp_h_blocks[m_comp_list[0]]; + m_mcus_per_col = m_comp_v_blocks[m_comp_list[0]]; + } + else + { + m_mcus_per_row = (((m_image_x_size + 7) / 8) + (max_h_samp - 1)) / max_h_samp; + m_mcus_per_col = (((m_image_y_size + 7) / 8) + (max_v_samp - 1)) / max_v_samp; + } + + if (m_comps_in_scan == 1) + { + m_mcu_org[0] = m_comp_list[0]; + + m_blocks_per_mcu = 1; + } + else + { + m_blocks_per_mcu = 0; + + for (component_num = 0; component_num < m_comps_in_scan; component_num++) + { + int num_blocks; + + component_id = m_comp_list[component_num]; + + num_blocks = m_comp_h_samp[component_id] * m_comp_v_samp[component_id]; + + while (num_blocks--) + m_mcu_org[m_blocks_per_mcu++] = component_id; + } + } + + if (m_blocks_per_mcu > m_max_blocks_per_mcu) + return false; + + for (int mcu_block = 0; mcu_block < m_blocks_per_mcu; mcu_block++) + { + int comp_id = m_mcu_org[mcu_block]; + if (comp_id >= JPGD_MAX_QUANT_TABLES) + return false; + } + + return true; + } + + // Starts a new scan. + int jpeg_decoder::init_scan() + { + if (!locate_sos_marker()) + return JPGD_FALSE; + + if (!calc_mcu_block_order()) + return JPGD_FALSE; + + check_huff_tables(); + + check_quant_tables(); + + memset(m_last_dc_val, 0, m_comps_in_frame * sizeof(uint)); + + m_eob_run = 0; + + if (m_restart_interval) + { + m_restarts_left = m_restart_interval; + m_next_restart_num = 0; + } + + fix_in_buffer(); + + return JPGD_TRUE; + } + + // Starts a frame. Determines if the number of components or sampling factors + // are supported. + void jpeg_decoder::init_frame() + { + int i; + + if (m_comps_in_frame == 1) + { + if ((m_comp_h_samp[0] != 1) || (m_comp_v_samp[0] != 1)) + stop_decoding(JPGD_UNSUPPORTED_SAMP_FACTORS); + + m_scan_type = JPGD_GRAYSCALE; + m_max_blocks_per_mcu = 1; + m_max_mcu_x_size = 8; + m_max_mcu_y_size = 8; + } + else if (m_comps_in_frame == 3) + { + if (((m_comp_h_samp[1] != 1) || (m_comp_v_samp[1] != 1)) || + ((m_comp_h_samp[2] != 1) || (m_comp_v_samp[2] != 1))) + stop_decoding(JPGD_UNSUPPORTED_SAMP_FACTORS); + + if ((m_comp_h_samp[0] == 1) && (m_comp_v_samp[0] == 1)) + { + m_scan_type = JPGD_YH1V1; + + m_max_blocks_per_mcu = 3; + m_max_mcu_x_size = 8; + m_max_mcu_y_size = 8; + } + else if ((m_comp_h_samp[0] == 2) && (m_comp_v_samp[0] == 1)) + { + m_scan_type = JPGD_YH2V1; + m_max_blocks_per_mcu = 4; + m_max_mcu_x_size = 16; + m_max_mcu_y_size = 8; + } + else if ((m_comp_h_samp[0] == 1) && (m_comp_v_samp[0] == 2)) + { + m_scan_type = JPGD_YH1V2; + m_max_blocks_per_mcu = 4; + m_max_mcu_x_size = 8; + m_max_mcu_y_size = 16; + } + else if ((m_comp_h_samp[0] == 2) && (m_comp_v_samp[0] == 2)) + { + m_scan_type = JPGD_YH2V2; + m_max_blocks_per_mcu = 6; + m_max_mcu_x_size = 16; + m_max_mcu_y_size = 16; + } + else + stop_decoding(JPGD_UNSUPPORTED_SAMP_FACTORS); + } + else + stop_decoding(JPGD_UNSUPPORTED_COLORSPACE); + + m_max_mcus_per_row = (m_image_x_size + (m_max_mcu_x_size - 1)) / m_max_mcu_x_size; + m_max_mcus_per_col = (m_image_y_size + (m_max_mcu_y_size - 1)) / m_max_mcu_y_size; + + // These values are for the *destination* pixels: after conversion. + if (m_scan_type == JPGD_GRAYSCALE) + m_dest_bytes_per_pixel = 1; + else + m_dest_bytes_per_pixel = 4; + + m_dest_bytes_per_scan_line = ((m_image_x_size + 15) & 0xFFF0) * m_dest_bytes_per_pixel; + + m_real_dest_bytes_per_scan_line = (m_image_x_size * m_dest_bytes_per_pixel); + + // Initialize two scan line buffers. + m_pScan_line_0 = (uint8*)alloc(m_dest_bytes_per_scan_line, true); + if ((m_scan_type == JPGD_YH1V2) || (m_scan_type == JPGD_YH2V2)) + m_pScan_line_1 = (uint8*)alloc(m_dest_bytes_per_scan_line, true); + + m_max_blocks_per_row = m_max_mcus_per_row * m_max_blocks_per_mcu; + + // Should never happen + if (m_max_blocks_per_row > JPGD_MAX_BLOCKS_PER_ROW) + stop_decoding(JPGD_DECODE_ERROR); + + // Allocate the coefficient buffer, enough for one MCU + m_pMCU_coefficients = (jpgd_block_t*)alloc(m_max_blocks_per_mcu * 64 * sizeof(jpgd_block_t)); + + for (i = 0; i < m_max_blocks_per_mcu; i++) + m_mcu_block_max_zag[i] = 64; + + m_pSample_buf = (uint8*)alloc(m_max_blocks_per_row * 64); + m_pSample_buf_prev = (uint8*)alloc(m_max_blocks_per_row * 64); + + m_total_lines_left = m_image_y_size; + + m_mcu_lines_left = 0; + + create_look_ups(); + } + + // The coeff_buf series of methods originally stored the coefficients + // into a "virtual" file which was located in EMS, XMS, or a disk file. A cache + // was used to make this process more efficient. Now, we can store the entire + // thing in RAM. + jpeg_decoder::coeff_buf* jpeg_decoder::coeff_buf_open(int block_num_x, int block_num_y, int block_len_x, int block_len_y) + { + coeff_buf* cb = (coeff_buf*)alloc(sizeof(coeff_buf)); + + cb->block_num_x = block_num_x; + cb->block_num_y = block_num_y; + cb->block_len_x = block_len_x; + cb->block_len_y = block_len_y; + cb->block_size = (block_len_x * block_len_y) * sizeof(jpgd_block_t); + cb->pData = (uint8*)alloc(cb->block_size * block_num_x * block_num_y, true); + return cb; + } + + inline jpgd_block_t* jpeg_decoder::coeff_buf_getp(coeff_buf* cb, int block_x, int block_y) + { + if ((block_x >= cb->block_num_x) || (block_y >= cb->block_num_y)) + stop_decoding(JPGD_DECODE_ERROR); + + return (jpgd_block_t*)(cb->pData + block_x * cb->block_size + block_y * (cb->block_size * cb->block_num_x)); + } + + // The following methods decode the various types of m_blocks encountered + // in progressively encoded images. + void jpeg_decoder::decode_block_dc_first(jpeg_decoder* pD, int component_id, int block_x, int block_y) + { + int s, r; + jpgd_block_t* p = pD->coeff_buf_getp(pD->m_dc_coeffs[component_id], block_x, block_y); + + if ((s = pD->huff_decode(pD->m_pHuff_tabs[pD->m_comp_dc_tab[component_id]])) != 0) + { + if (s >= 16) + pD->stop_decoding(JPGD_DECODE_ERROR); + + r = pD->get_bits_no_markers(s); + s = JPGD_HUFF_EXTEND(r, s); + } + + pD->m_last_dc_val[component_id] = (s += pD->m_last_dc_val[component_id]); + + p[0] = static_cast(s << pD->m_successive_low); + } + + void jpeg_decoder::decode_block_dc_refine(jpeg_decoder* pD, int component_id, int block_x, int block_y) + { + if (pD->get_bits_no_markers(1)) + { + jpgd_block_t* p = pD->coeff_buf_getp(pD->m_dc_coeffs[component_id], block_x, block_y); + + p[0] |= (1 << pD->m_successive_low); + } + } + + void jpeg_decoder::decode_block_ac_first(jpeg_decoder* pD, int component_id, int block_x, int block_y) + { + int k, s, r; + + if (pD->m_eob_run) + { + pD->m_eob_run--; + return; + } + + jpgd_block_t* p = pD->coeff_buf_getp(pD->m_ac_coeffs[component_id], block_x, block_y); + + for (k = pD->m_spectral_start; k <= pD->m_spectral_end; k++) + { + unsigned int idx = pD->m_comp_ac_tab[component_id]; + if (idx >= JPGD_MAX_HUFF_TABLES) + pD->stop_decoding(JPGD_DECODE_ERROR); + + s = pD->huff_decode(pD->m_pHuff_tabs[idx]); + + r = s >> 4; + s &= 15; + + if (s) + { + if ((k += r) > 63) + pD->stop_decoding(JPGD_DECODE_ERROR); + + r = pD->get_bits_no_markers(s); + s = JPGD_HUFF_EXTEND(r, s); + + p[g_ZAG[k]] = static_cast(s << pD->m_successive_low); + } + else + { + if (r == 15) + { + if ((k += 15) > 63) + pD->stop_decoding(JPGD_DECODE_ERROR); + } + else + { + pD->m_eob_run = 1 << r; + + if (r) + pD->m_eob_run += pD->get_bits_no_markers(r); + + pD->m_eob_run--; + + break; + } + } + } + } + + void jpeg_decoder::decode_block_ac_refine(jpeg_decoder* pD, int component_id, int block_x, int block_y) + { + int s, k, r; + + int p1 = 1 << pD->m_successive_low; + + //int m1 = (-1) << pD->m_successive_low; + int m1 = static_cast((UINT32_MAX << pD->m_successive_low)); + + jpgd_block_t* p = pD->coeff_buf_getp(pD->m_ac_coeffs[component_id], block_x, block_y); + if (pD->m_spectral_end > 63) + pD->stop_decoding(JPGD_DECODE_ERROR); + + k = pD->m_spectral_start; + + if (pD->m_eob_run == 0) + { + for (; k <= pD->m_spectral_end; k++) + { + unsigned int idx = pD->m_comp_ac_tab[component_id]; + if (idx >= JPGD_MAX_HUFF_TABLES) + pD->stop_decoding(JPGD_DECODE_ERROR); + + s = pD->huff_decode(pD->m_pHuff_tabs[idx]); + + r = s >> 4; + s &= 15; + + if (s) + { + if (s != 1) + pD->stop_decoding(JPGD_DECODE_ERROR); + + if (pD->get_bits_no_markers(1)) + s = p1; + else + s = m1; + } + else + { + if (r != 15) + { + pD->m_eob_run = 1 << r; + + if (r) + pD->m_eob_run += pD->get_bits_no_markers(r); + + break; + } + } + + do + { + jpgd_block_t* this_coef = p + g_ZAG[k & 63]; + + if (*this_coef != 0) + { + if (pD->get_bits_no_markers(1)) + { + if ((*this_coef & p1) == 0) + { + if (*this_coef >= 0) + *this_coef = static_cast(*this_coef + p1); + else + *this_coef = static_cast(*this_coef + m1); + } + } + } + else + { + if (--r < 0) + break; + } + + k++; + + } while (k <= pD->m_spectral_end); + + if ((s) && (k < 64)) + { + p[g_ZAG[k]] = static_cast(s); + } + } + } + + if (pD->m_eob_run > 0) + { + for (; k <= pD->m_spectral_end; k++) + { + jpgd_block_t* this_coef = p + g_ZAG[k & 63]; // logical AND to shut up static code analysis + + if (*this_coef != 0) + { + if (pD->get_bits_no_markers(1)) + { + if ((*this_coef & p1) == 0) + { + if (*this_coef >= 0) + *this_coef = static_cast(*this_coef + p1); + else + *this_coef = static_cast(*this_coef + m1); + } + } + } + } + + pD->m_eob_run--; + } + } + + // Decode a scan in a progressively encoded image. + void jpeg_decoder::decode_scan(pDecode_block_func decode_block_func) + { + int mcu_row, mcu_col, mcu_block; + int block_x_mcu[JPGD_MAX_COMPONENTS], block_y_mcu[JPGD_MAX_COMPONENTS]; + + memset(block_y_mcu, 0, sizeof(block_y_mcu)); + + for (mcu_col = 0; mcu_col < m_mcus_per_col; mcu_col++) + { + int component_num, component_id; + + memset(block_x_mcu, 0, sizeof(block_x_mcu)); + + for (mcu_row = 0; mcu_row < m_mcus_per_row; mcu_row++) + { + int block_x_mcu_ofs = 0, block_y_mcu_ofs = 0; + + if ((m_restart_interval) && (m_restarts_left == 0)) + process_restart(); + + for (mcu_block = 0; mcu_block < m_blocks_per_mcu; mcu_block++) + { + component_id = m_mcu_org[mcu_block]; + + decode_block_func(this, component_id, block_x_mcu[component_id] + block_x_mcu_ofs, block_y_mcu[component_id] + block_y_mcu_ofs); + + if (m_comps_in_scan == 1) + block_x_mcu[component_id]++; + else + { + if (++block_x_mcu_ofs == m_comp_h_samp[component_id]) + { + block_x_mcu_ofs = 0; + + if (++block_y_mcu_ofs == m_comp_v_samp[component_id]) + { + block_y_mcu_ofs = 0; + block_x_mcu[component_id] += m_comp_h_samp[component_id]; + } + } + } + } + + m_restarts_left--; + } + + if (m_comps_in_scan == 1) + block_y_mcu[m_comp_list[0]]++; + else + { + for (component_num = 0; component_num < m_comps_in_scan; component_num++) + { + component_id = m_comp_list[component_num]; + block_y_mcu[component_id] += m_comp_v_samp[component_id]; + } + } + } + } + + // Decode a progressively encoded image. + void jpeg_decoder::init_progressive() + { + int i; + + if (m_comps_in_frame == 4) + stop_decoding(JPGD_UNSUPPORTED_COLORSPACE); + + // Allocate the coefficient buffers. + for (i = 0; i < m_comps_in_frame; i++) + { + m_dc_coeffs[i] = coeff_buf_open(m_max_mcus_per_row * m_comp_h_samp[i], m_max_mcus_per_col * m_comp_v_samp[i], 1, 1); + m_ac_coeffs[i] = coeff_buf_open(m_max_mcus_per_row * m_comp_h_samp[i], m_max_mcus_per_col * m_comp_v_samp[i], 8, 8); + } + + // See https://libjpeg-turbo.org/pmwiki/uploads/About/TwoIssueswiththeJPEGStandard.pdf + uint32_t total_scans = 0; + const uint32_t MAX_SCANS_TO_PROCESS = 1000; + + for (; ; ) + { + int dc_only_scan, refinement_scan; + pDecode_block_func decode_block_func; + + if (!init_scan()) + break; + + dc_only_scan = (m_spectral_start == 0); + refinement_scan = (m_successive_high != 0); + + if ((m_spectral_start > m_spectral_end) || (m_spectral_end > 63)) + stop_decoding(JPGD_BAD_SOS_SPECTRAL); + + if (dc_only_scan) + { + if (m_spectral_end) + stop_decoding(JPGD_BAD_SOS_SPECTRAL); + } + else if (m_comps_in_scan != 1) /* AC scans can only contain one component */ + stop_decoding(JPGD_BAD_SOS_SPECTRAL); + + if ((refinement_scan) && (m_successive_low != m_successive_high - 1)) + stop_decoding(JPGD_BAD_SOS_SUCCESSIVE); + + if (dc_only_scan) + { + if (refinement_scan) + decode_block_func = decode_block_dc_refine; + else + decode_block_func = decode_block_dc_first; + } + else + { + if (refinement_scan) + decode_block_func = decode_block_ac_refine; + else + decode_block_func = decode_block_ac_first; + } + + decode_scan(decode_block_func); + + m_bits_left = 16; + get_bits(16); + get_bits(16); + + total_scans++; + if (total_scans > MAX_SCANS_TO_PROCESS) + stop_decoding(JPGD_TOO_MANY_SCANS); + } + + m_comps_in_scan = m_comps_in_frame; + + for (i = 0; i < m_comps_in_frame; i++) + m_comp_list[i] = i; + + if (!calc_mcu_block_order()) + stop_decoding(JPGD_DECODE_ERROR); + } + + void jpeg_decoder::init_sequential() + { + if (!init_scan()) + stop_decoding(JPGD_UNEXPECTED_MARKER); + } + + void jpeg_decoder::decode_start() + { + init_frame(); + + if (m_progressive_flag) + init_progressive(); + else + init_sequential(); + } + + void jpeg_decoder::decode_init(jpeg_decoder_stream* pStream, uint32_t flags) + { + init(pStream, flags); + locate_sof_marker(); + } + + jpeg_decoder::jpeg_decoder(jpeg_decoder_stream* pStream, uint32_t flags) + { + if (setjmp(m_jmp_state)) + return; + decode_init(pStream, flags); + } + + int jpeg_decoder::begin_decoding() + { + if (m_ready_flag) + return JPGD_SUCCESS; + + if (m_error_code) + return JPGD_FAILED; + + if (setjmp(m_jmp_state)) + return JPGD_FAILED; + + decode_start(); + + m_ready_flag = true; + + return JPGD_SUCCESS; + } + + jpeg_decoder::~jpeg_decoder() + { + free_all_blocks(); + } + + jpeg_decoder_file_stream::jpeg_decoder_file_stream() + { + m_pFile = nullptr; + m_eof_flag = false; + m_error_flag = false; + } + + void jpeg_decoder_file_stream::close() + { + if (m_pFile) + { + fclose(m_pFile); + m_pFile = nullptr; + } + + m_eof_flag = false; + m_error_flag = false; + } + + jpeg_decoder_file_stream::~jpeg_decoder_file_stream() + { + close(); + } + + bool jpeg_decoder_file_stream::open(const char* Pfilename) + { + close(); + + m_eof_flag = false; + m_error_flag = false; + +#if defined(_MSC_VER) + m_pFile = nullptr; + fopen_s(&m_pFile, Pfilename, "rb"); +#else + m_pFile = fopen(Pfilename, "rb"); +#endif + return m_pFile != nullptr; + } + + int jpeg_decoder_file_stream::read(uint8* pBuf, int max_bytes_to_read, bool* pEOF_flag) + { + if (!m_pFile) + return -1; + + if (m_eof_flag) + { + *pEOF_flag = true; + return 0; + } + + if (m_error_flag) + return -1; + + int bytes_read = static_cast(fread(pBuf, 1, max_bytes_to_read, m_pFile)); + if (bytes_read < max_bytes_to_read) + { + if (ferror(m_pFile)) + { + m_error_flag = true; + return -1; + } + + m_eof_flag = true; + *pEOF_flag = true; + } + + return bytes_read; + } + + bool jpeg_decoder_mem_stream::open(const uint8* pSrc_data, uint size) + { + close(); + m_pSrc_data = pSrc_data; + m_ofs = 0; + m_size = size; + return true; + } + + int jpeg_decoder_mem_stream::read(uint8* pBuf, int max_bytes_to_read, bool* pEOF_flag) + { + *pEOF_flag = false; + + if (!m_pSrc_data) + return -1; + + uint bytes_remaining = m_size - m_ofs; + if ((uint)max_bytes_to_read > bytes_remaining) + { + max_bytes_to_read = bytes_remaining; + *pEOF_flag = true; + } + + memcpy(pBuf, m_pSrc_data + m_ofs, max_bytes_to_read); + m_ofs += max_bytes_to_read; + + return max_bytes_to_read; + } + + unsigned char* decompress_jpeg_image_from_stream(jpeg_decoder_stream* pStream, int* width, int* height, int* actual_comps, int req_comps, uint32_t flags) + { + if (!actual_comps) + return nullptr; + *actual_comps = 0; + + if ((!pStream) || (!width) || (!height) || (!req_comps)) + return nullptr; + + if ((req_comps != 1) && (req_comps != 3) && (req_comps != 4)) + return nullptr; + + jpeg_decoder decoder(pStream, flags); + if (decoder.get_error_code() != JPGD_SUCCESS) + return nullptr; + + const int image_width = decoder.get_width(), image_height = decoder.get_height(); + *width = image_width; + *height = image_height; + *actual_comps = decoder.get_num_components(); + + if (decoder.begin_decoding() != JPGD_SUCCESS) + return nullptr; + + const int dst_bpl = image_width * req_comps; + + uint8* pImage_data = (uint8*)jpgd_malloc(dst_bpl * image_height); + if (!pImage_data) + return nullptr; + + for (int y = 0; y < image_height; y++) + { + const uint8* pScan_line = nullptr; + uint scan_line_len; + if (decoder.decode((const void**)&pScan_line, &scan_line_len) != JPGD_SUCCESS) + { + jpgd_free(pImage_data); + return nullptr; + } + + uint8* pDst = pImage_data + y * dst_bpl; + + if (((req_comps == 1) && (decoder.get_num_components() == 1)) || ((req_comps == 4) && (decoder.get_num_components() == 3))) + memcpy(pDst, pScan_line, dst_bpl); + else if (decoder.get_num_components() == 1) + { + if (req_comps == 3) + { + for (int x = 0; x < image_width; x++) + { + uint8 luma = pScan_line[x]; + pDst[0] = luma; + pDst[1] = luma; + pDst[2] = luma; + pDst += 3; + } + } + else + { + for (int x = 0; x < image_width; x++) + { + uint8 luma = pScan_line[x]; + pDst[0] = luma; + pDst[1] = luma; + pDst[2] = luma; + pDst[3] = 255; + pDst += 4; + } + } + } + else if (decoder.get_num_components() == 3) + { + if (req_comps == 1) + { + const int YR = 19595, YG = 38470, YB = 7471; + for (int x = 0; x < image_width; x++) + { + int r = pScan_line[x * 4 + 0]; + int g = pScan_line[x * 4 + 1]; + int b = pScan_line[x * 4 + 2]; + *pDst++ = static_cast((r * YR + g * YG + b * YB + 32768) >> 16); + } + } + else + { + for (int x = 0; x < image_width; x++) + { + pDst[0] = pScan_line[x * 4 + 0]; + pDst[1] = pScan_line[x * 4 + 1]; + pDst[2] = pScan_line[x * 4 + 2]; + pDst += 3; + } + } + } + } + + return pImage_data; + } + + unsigned char* decompress_jpeg_image_from_memory(const unsigned char* pSrc_data, int src_data_size, int* width, int* height, int* actual_comps, int req_comps, uint32_t flags) + { + jpgd::jpeg_decoder_mem_stream mem_stream(pSrc_data, src_data_size); + return decompress_jpeg_image_from_stream(&mem_stream, width, height, actual_comps, req_comps, flags); + } + + unsigned char* decompress_jpeg_image_from_file(const char* pSrc_filename, int* width, int* height, int* actual_comps, int req_comps, uint32_t flags) + { + jpgd::jpeg_decoder_file_stream file_stream; + if (!file_stream.open(pSrc_filename)) + return nullptr; + return decompress_jpeg_image_from_stream(&file_stream, width, height, actual_comps, req_comps, flags); + } + +} // namespace jpgd diff --git a/tools/imageio/external/jpgd.h b/tools/imageio/external/jpgd.h new file mode 100644 index 0000000000..92e53335c6 --- /dev/null +++ b/tools/imageio/external/jpgd.h @@ -0,0 +1,347 @@ +// jpgd.h - C++ class for JPEG decompression. +// Public domain, Rich Geldreich +#ifndef JPEG_DECODER_H +#define JPEG_DECODER_H + +#include +#include +#include +#include +#include + +#ifdef _MSC_VER +#define JPGD_NORETURN __declspec(noreturn) +#elif defined(__GNUC__) +#define JPGD_NORETURN __attribute__ ((noreturn)) +#else +#define JPGD_NORETURN +#endif + +#define JPGD_HUFF_TREE_MAX_LENGTH 512 +#define JPGD_HUFF_CODE_SIZE_MAX_LENGTH 256 + +namespace jpgd +{ + typedef unsigned char uint8; + typedef signed short int16; + typedef unsigned short uint16; + typedef unsigned int uint; + typedef signed int int32; + + // Loads a JPEG image from a memory buffer or a file. + // req_comps can be 1 (grayscale), 3 (RGB), or 4 (RGBA). + // On return, width/height will be set to the image's dimensions, and actual_comps will be set to the either 1 (grayscale) or 3 (RGB). + // Notes: For more control over where and how the source data is read, see the decompress_jpeg_image_from_stream() function below, or call the jpeg_decoder class directly. + // Requesting a 8 or 32bpp image is currently a little faster than 24bpp because the jpeg_decoder class itself currently always unpacks to either 8 or 32bpp. + unsigned char* decompress_jpeg_image_from_memory(const unsigned char* pSrc_data, int src_data_size, int* width, int* height, int* actual_comps, int req_comps, uint32_t flags = 0); + unsigned char* decompress_jpeg_image_from_file(const char* pSrc_filename, int* width, int* height, int* actual_comps, int req_comps, uint32_t flags = 0); + + // Success/failure error codes. + enum jpgd_status + { + JPGD_SUCCESS = 0, JPGD_FAILED = -1, JPGD_DONE = 1, + JPGD_BAD_DHT_COUNTS = -256, JPGD_BAD_DHT_INDEX, JPGD_BAD_DHT_MARKER, JPGD_BAD_DQT_MARKER, JPGD_BAD_DQT_TABLE, + JPGD_BAD_PRECISION, JPGD_BAD_HEIGHT, JPGD_BAD_WIDTH, JPGD_TOO_MANY_COMPONENTS, + JPGD_BAD_SOF_LENGTH, JPGD_BAD_VARIABLE_MARKER, JPGD_BAD_DRI_LENGTH, JPGD_BAD_SOS_LENGTH, + JPGD_BAD_SOS_COMP_ID, JPGD_W_EXTRA_BYTES_BEFORE_MARKER, JPGD_NO_ARITHMITIC_SUPPORT, JPGD_UNEXPECTED_MARKER, + JPGD_NOT_JPEG, JPGD_UNSUPPORTED_MARKER, JPGD_BAD_DQT_LENGTH, JPGD_TOO_MANY_BLOCKS, + JPGD_UNDEFINED_QUANT_TABLE, JPGD_UNDEFINED_HUFF_TABLE, JPGD_NOT_SINGLE_SCAN, JPGD_UNSUPPORTED_COLORSPACE, + JPGD_UNSUPPORTED_SAMP_FACTORS, JPGD_DECODE_ERROR, JPGD_BAD_RESTART_MARKER, + JPGD_BAD_SOS_SPECTRAL, JPGD_BAD_SOS_SUCCESSIVE, JPGD_STREAM_READ, JPGD_NOTENOUGHMEM, JPGD_TOO_MANY_SCANS + }; + + // Input stream interface. + // Derive from this class to read input data from sources other than files or memory. Set m_eof_flag to true when no more data is available. + // The decoder is rather greedy: it will keep on calling this method until its internal input buffer is full, or until the EOF flag is set. + // It the input stream contains data after the JPEG stream's EOI (end of image) marker it will probably be pulled into the internal buffer. + // Call the get_total_bytes_read() method to determine the actual size of the JPEG stream after successful decoding. + class jpeg_decoder_stream + { + public: + jpeg_decoder_stream() { } + virtual ~jpeg_decoder_stream() { } + + // The read() method is called when the internal input buffer is empty. + // Parameters: + // pBuf - input buffer + // max_bytes_to_read - maximum bytes that can be written to pBuf + // pEOF_flag - set this to true if at end of stream (no more bytes remaining) + // Returns -1 on error, otherwise return the number of bytes actually written to the buffer (which may be 0). + // Notes: This method will be called in a loop until you set *pEOF_flag to true or the internal buffer is full. + virtual int read(uint8* pBuf, int max_bytes_to_read, bool* pEOF_flag) = 0; + }; + + // stdio FILE stream class. + class jpeg_decoder_file_stream : public jpeg_decoder_stream + { + jpeg_decoder_file_stream(const jpeg_decoder_file_stream&); + jpeg_decoder_file_stream& operator =(const jpeg_decoder_file_stream&); + + FILE* m_pFile; + bool m_eof_flag, m_error_flag; + + public: + jpeg_decoder_file_stream(); + virtual ~jpeg_decoder_file_stream(); + + bool open(const char* Pfilename); + void close(); + + virtual int read(uint8* pBuf, int max_bytes_to_read, bool* pEOF_flag); + }; + + // Memory stream class. + class jpeg_decoder_mem_stream : public jpeg_decoder_stream + { + const uint8* m_pSrc_data; + uint m_ofs, m_size; + + public: + jpeg_decoder_mem_stream() : m_pSrc_data(NULL), m_ofs(0), m_size(0) { } + jpeg_decoder_mem_stream(const uint8* pSrc_data, uint size) : m_pSrc_data(pSrc_data), m_ofs(0), m_size(size) { } + + virtual ~jpeg_decoder_mem_stream() { } + + bool open(const uint8* pSrc_data, uint size); + void close() { m_pSrc_data = NULL; m_ofs = 0; m_size = 0; } + + virtual int read(uint8* pBuf, int max_bytes_to_read, bool* pEOF_flag); + }; + + // Loads JPEG file from a jpeg_decoder_stream. + unsigned char* decompress_jpeg_image_from_stream(jpeg_decoder_stream* pStream, int* width, int* height, int* actual_comps, int req_comps, uint32_t flags = 0); + + enum + { + JPGD_IN_BUF_SIZE = 8192, JPGD_MAX_BLOCKS_PER_MCU = 10, JPGD_MAX_HUFF_TABLES = 8, JPGD_MAX_QUANT_TABLES = 4, + JPGD_MAX_COMPONENTS = 4, JPGD_MAX_COMPS_IN_SCAN = 4, JPGD_MAX_BLOCKS_PER_ROW = 16384, JPGD_MAX_HEIGHT = 32768, JPGD_MAX_WIDTH = 32768 + }; + + typedef int16 jpgd_quant_t; + typedef int16 jpgd_block_t; + + class jpeg_decoder + { + public: + enum + { + cFlagLinearChromaFiltering = 1 + }; + + // Call get_error_code() after constructing to determine if the stream is valid or not. You may call the get_width(), get_height(), etc. + // methods after the constructor is called. You may then either destruct the object, or begin decoding the image by calling begin_decoding(), then decode() on each scanline. + jpeg_decoder(jpeg_decoder_stream* pStream, uint32_t flags = cFlagLinearChromaFiltering); + + ~jpeg_decoder(); + + // Call this method after constructing the object to begin decompression. + // If JPGD_SUCCESS is returned you may then call decode() on each scanline. + + int begin_decoding(); + + // Returns the next scan line. + // For grayscale images, pScan_line will point to a buffer containing 8-bit pixels (get_bytes_per_pixel() will return 1). + // Otherwise, it will always point to a buffer containing 32-bit RGBA pixels (A will always be 255, and get_bytes_per_pixel() will return 4). + // Returns JPGD_SUCCESS if a scan line has been returned. + // Returns JPGD_DONE if all scan lines have been returned. + // Returns JPGD_FAILED if an error occurred. Call get_error_code() for a more info. + int decode(const void** pScan_line, uint* pScan_line_len); + + inline jpgd_status get_error_code() const { return m_error_code; } + + inline int get_width() const { return m_image_x_size; } + inline int get_height() const { return m_image_y_size; } + + inline int get_num_components() const { return m_comps_in_frame; } + + inline int get_bytes_per_pixel() const { return m_dest_bytes_per_pixel; } + inline int get_bytes_per_scan_line() const { return m_image_x_size * get_bytes_per_pixel(); } + + // Returns the total number of bytes actually consumed by the decoder (which should equal the actual size of the JPEG file). + inline int get_total_bytes_read() const { return m_total_bytes_read; } + + private: + jpeg_decoder(const jpeg_decoder&); + jpeg_decoder& operator =(const jpeg_decoder&); + + typedef void (*pDecode_block_func)(jpeg_decoder*, int, int, int); + + struct huff_tables + { + bool ac_table; + uint look_up[256]; + uint look_up2[256]; + uint8 code_size[JPGD_HUFF_CODE_SIZE_MAX_LENGTH]; + uint tree[JPGD_HUFF_TREE_MAX_LENGTH]; + }; + + struct coeff_buf + { + uint8* pData; + int block_num_x, block_num_y; + int block_len_x, block_len_y; + int block_size; + }; + + struct mem_block + { + mem_block* m_pNext; + size_t m_used_count; + size_t m_size; + char m_data[1]; + }; + + jmp_buf m_jmp_state; + uint32_t m_flags; + mem_block* m_pMem_blocks; + int m_image_x_size; + int m_image_y_size; + jpeg_decoder_stream* m_pStream; + + int m_progressive_flag; + + uint8 m_huff_ac[JPGD_MAX_HUFF_TABLES]; + uint8* m_huff_num[JPGD_MAX_HUFF_TABLES]; // pointer to number of Huffman codes per bit size + uint8* m_huff_val[JPGD_MAX_HUFF_TABLES]; // pointer to Huffman codes per bit size + jpgd_quant_t* m_quant[JPGD_MAX_QUANT_TABLES]; // pointer to quantization tables + int m_scan_type; // Gray, Yh1v1, Yh1v2, Yh2v1, Yh2v2 (CMYK111, CMYK4114 no longer supported) + int m_comps_in_frame; // # of components in frame + int m_comp_h_samp[JPGD_MAX_COMPONENTS]; // component's horizontal sampling factor + int m_comp_v_samp[JPGD_MAX_COMPONENTS]; // component's vertical sampling factor + int m_comp_quant[JPGD_MAX_COMPONENTS]; // component's quantization table selector + int m_comp_ident[JPGD_MAX_COMPONENTS]; // component's ID + int m_comp_h_blocks[JPGD_MAX_COMPONENTS]; + int m_comp_v_blocks[JPGD_MAX_COMPONENTS]; + int m_comps_in_scan; // # of components in scan + int m_comp_list[JPGD_MAX_COMPS_IN_SCAN]; // components in this scan + int m_comp_dc_tab[JPGD_MAX_COMPONENTS]; // component's DC Huffman coding table selector + int m_comp_ac_tab[JPGD_MAX_COMPONENTS]; // component's AC Huffman coding table selector + int m_spectral_start; // spectral selection start + int m_spectral_end; // spectral selection end + int m_successive_low; // successive approximation low + int m_successive_high; // successive approximation high + int m_max_mcu_x_size; // MCU's max. X size in pixels + int m_max_mcu_y_size; // MCU's max. Y size in pixels + int m_blocks_per_mcu; + int m_max_blocks_per_row; + int m_mcus_per_row, m_mcus_per_col; + int m_mcu_org[JPGD_MAX_BLOCKS_PER_MCU]; + int m_total_lines_left; // total # lines left in image + int m_mcu_lines_left; // total # lines left in this MCU + int m_num_buffered_scanlines; + int m_real_dest_bytes_per_scan_line; + int m_dest_bytes_per_scan_line; // rounded up + int m_dest_bytes_per_pixel; // 4 (RGB) or 1 (Y) + huff_tables* m_pHuff_tabs[JPGD_MAX_HUFF_TABLES]; + coeff_buf* m_dc_coeffs[JPGD_MAX_COMPONENTS]; + coeff_buf* m_ac_coeffs[JPGD_MAX_COMPONENTS]; + int m_eob_run; + int m_block_y_mcu[JPGD_MAX_COMPONENTS]; + uint8* m_pIn_buf_ofs; + int m_in_buf_left; + int m_tem_flag; + + uint8 m_in_buf_pad_start[64]; + uint8 m_in_buf[JPGD_IN_BUF_SIZE + 128]; + uint8 m_in_buf_pad_end[64]; + + int m_bits_left; + uint m_bit_buf; + int m_restart_interval; + int m_restarts_left; + int m_next_restart_num; + int m_max_mcus_per_row; + int m_max_blocks_per_mcu; + + int m_max_mcus_per_col; + uint m_last_dc_val[JPGD_MAX_COMPONENTS]; + jpgd_block_t* m_pMCU_coefficients; + int m_mcu_block_max_zag[JPGD_MAX_BLOCKS_PER_MCU]; + uint8* m_pSample_buf; + uint8* m_pSample_buf_prev; + int m_crr[256]; + int m_cbb[256]; + int m_crg[256]; + int m_cbg[256]; + uint8* m_pScan_line_0; + uint8* m_pScan_line_1; + jpgd_status m_error_code; + int m_total_bytes_read; + + bool m_ready_flag; + bool m_eof_flag; + bool m_sample_buf_prev_valid; + + inline int check_sample_buf_ofs(int ofs) const { assert(ofs >= 0); assert(ofs < m_max_blocks_per_row * 64); return ofs; } + void free_all_blocks(); + JPGD_NORETURN void stop_decoding(jpgd_status status); + void* alloc(size_t n, bool zero = false); + void word_clear(void* p, uint16 c, uint n); + void prep_in_buffer(); + void read_dht_marker(); + void read_dqt_marker(); + void read_sof_marker(); + void skip_variable_marker(); + void read_dri_marker(); + void read_sos_marker(); + int next_marker(); + int process_markers(); + void locate_soi_marker(); + void locate_sof_marker(); + int locate_sos_marker(); + void init(jpeg_decoder_stream* pStream, uint32_t flags); + void create_look_ups(); + void fix_in_buffer(); + void transform_mcu(int mcu_row); + coeff_buf* coeff_buf_open(int block_num_x, int block_num_y, int block_len_x, int block_len_y); + inline jpgd_block_t* coeff_buf_getp(coeff_buf* cb, int block_x, int block_y); + void load_next_row(); + void decode_next_row(); + void make_huff_table(int index, huff_tables* pH); + void check_quant_tables(); + void check_huff_tables(); + bool calc_mcu_block_order(); + int init_scan(); + void init_frame(); + void process_restart(); + void decode_scan(pDecode_block_func decode_block_func); + void init_progressive(); + void init_sequential(); + void decode_start(); + void decode_init(jpeg_decoder_stream* pStream, uint32_t flags); + void H2V2Convert(); + uint32_t H2V2ConvertFiltered(); + void H2V1Convert(); + void H2V1ConvertFiltered(); + void H1V2Convert(); + void H1V2ConvertFiltered(); + void H1V1Convert(); + void gray_convert(); + void find_eoi(); + inline uint get_char(); + inline uint get_char(bool* pPadding_flag); + inline void stuff_char(uint8 q); + inline uint8 get_octet(); + inline uint get_bits(int num_bits); + inline uint get_bits_no_markers(int numbits); + inline int huff_decode(huff_tables* pH); + inline int huff_decode(huff_tables* pH, int& extrabits); + + // Clamps a value between 0-255. + static inline uint8 clamp(int i) + { + if (static_cast(i) > 255) + i = (((~i) >> 31) & 0xFF); + return static_cast(i); + } + int decode_next_mcu_row(); + + static void decode_block_dc_first(jpeg_decoder* pD, int component_id, int block_x, int block_y); + static void decode_block_dc_refine(jpeg_decoder* pD, int component_id, int block_x, int block_y); + static void decode_block_ac_first(jpeg_decoder* pD, int component_id, int block_x, int block_y); + static void decode_block_ac_refine(jpeg_decoder* pD, int component_id, int block_x, int block_y); + }; + +} // namespace jpgd + +#endif // JPEG_DECODER_H diff --git a/tools/imageio/external/tinyexr.h b/tools/imageio/external/tinyexr.h new file mode 100644 index 0000000000..6d14d48682 --- /dev/null +++ b/tools/imageio/external/tinyexr.h @@ -0,0 +1,13581 @@ +/* +Copyright (c) 2014 - 2019, Syoyo Fujita and many contributors. +All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + * Neither the name of the Syoyo Fujita nor the + names of its contributors may be used to endorse or promote products + derived from this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND +ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED +WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +DISCLAIMED. IN NO EVENT SHALL BE LIABLE FOR ANY +DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES +(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; +LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND +ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +*/ + +// TinyEXR contains some OpenEXR code, which is licensed under ------------ + +/////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2002, Industrial Light & Magic, a division of Lucas +// Digital Ltd. LLC +// +// All rights reserved. +// +// Redistribution and use in source and binary forms, with or without +// modification, are permitted provided that the following conditions are +// met: +// * Redistributions of source code must retain the above copyright +// notice, this list of conditions and the following disclaimer. +// * Redistributions in binary form must reproduce the above +// copyright notice, this list of conditions and the following disclaimer +// in the documentation and/or other materials provided with the +// distribution. +// * Neither the name of Industrial Light & Magic nor the names of +// its contributors may be used to endorse or promote products derived +// from this software without specific prior written permission. +// +// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +// +/////////////////////////////////////////////////////////////////////////// + +// End of OpenEXR license ------------------------------------------------- + +#ifndef TINYEXR_H_ +#define TINYEXR_H_ + +// +// +// Do this: +// #define TINYEXR_IMPLEMENTATION +// before you include this file in *one* C or C++ file to create the +// implementation. +// +// // i.e. it should look like this: +// #include ... +// #include ... +// #include ... +// #define TINYEXR_IMPLEMENTATION +// #include "tinyexr.h" +// +// + +#include // for size_t +#include // guess stdint.h is available(C99) + +#ifdef __cplusplus +extern "C" { +#endif + +// Use embedded miniz or not to decode ZIP format pixel. Linking with zlib +// required if this flas is 0. +#ifndef TINYEXR_USE_MINIZ +#define TINYEXR_USE_MINIZ (1) +#endif + +// Disable PIZ comporession when applying cpplint. +#ifndef TINYEXR_USE_PIZ +#define TINYEXR_USE_PIZ (1) +#endif + +#ifndef TINYEXR_USE_ZFP +#define TINYEXR_USE_ZFP (0) // TinyEXR extension. +// http://computation.llnl.gov/projects/floating-point-compression +#endif + +#ifndef TINYEXR_USE_THREAD +#define TINYEXR_USE_THREAD (0) // No threaded loading. +// http://computation.llnl.gov/projects/floating-point-compression +#endif + +#ifndef TINYEXR_USE_OPENMP +#ifdef _OPENMP +#define TINYEXR_USE_OPENMP (1) +#else +#define TINYEXR_USE_OPENMP (0) +#endif +#endif + +#define TINYEXR_SUCCESS (0) +#define TINYEXR_ERROR_INVALID_MAGIC_NUMBER (-1) +#define TINYEXR_ERROR_INVALID_EXR_VERSION (-2) +#define TINYEXR_ERROR_INVALID_ARGUMENT (-3) +#define TINYEXR_ERROR_INVALID_DATA (-4) +#define TINYEXR_ERROR_INVALID_FILE (-5) +#define TINYEXR_ERROR_INVALID_PARAMETER (-6) +#define TINYEXR_ERROR_CANT_OPEN_FILE (-7) +#define TINYEXR_ERROR_UNSUPPORTED_FORMAT (-8) +#define TINYEXR_ERROR_INVALID_HEADER (-9) +#define TINYEXR_ERROR_UNSUPPORTED_FEATURE (-10) +#define TINYEXR_ERROR_CANT_WRITE_FILE (-11) +#define TINYEXR_ERROR_SERIALZATION_FAILED (-12) +#define TINYEXR_ERROR_LAYER_NOT_FOUND (-13) + +// @note { OpenEXR file format: http://www.openexr.com/openexrfilelayout.pdf } + +// pixel type: possible values are: UINT = 0 HALF = 1 FLOAT = 2 +#define TINYEXR_PIXELTYPE_UINT (0) +#define TINYEXR_PIXELTYPE_HALF (1) +#define TINYEXR_PIXELTYPE_FLOAT (2) + +#define TINYEXR_MAX_HEADER_ATTRIBUTES (1024) +#define TINYEXR_MAX_CUSTOM_ATTRIBUTES (128) + +#define TINYEXR_COMPRESSIONTYPE_NONE (0) +#define TINYEXR_COMPRESSIONTYPE_RLE (1) +#define TINYEXR_COMPRESSIONTYPE_ZIPS (2) +#define TINYEXR_COMPRESSIONTYPE_ZIP (3) +#define TINYEXR_COMPRESSIONTYPE_PIZ (4) +#define TINYEXR_COMPRESSIONTYPE_ZFP (128) // TinyEXR extension + +#define TINYEXR_ZFP_COMPRESSIONTYPE_RATE (0) +#define TINYEXR_ZFP_COMPRESSIONTYPE_PRECISION (1) +#define TINYEXR_ZFP_COMPRESSIONTYPE_ACCURACY (2) + +#define TINYEXR_TILE_ONE_LEVEL (0) +#define TINYEXR_TILE_MIPMAP_LEVELS (1) +#define TINYEXR_TILE_RIPMAP_LEVELS (2) + +#define TINYEXR_TILE_ROUND_DOWN (0) +#define TINYEXR_TILE_ROUND_UP (1) + +typedef struct _EXRVersion { + int version; // this must be 2 + int tiled; // tile format image + int long_name; // long name attribute + int non_image; // deep image(EXR 2.0) + int multipart; // multi-part(EXR 2.0) +} EXRVersion; + +typedef struct _EXRAttribute { + char name[256]; // name and type are up to 255 chars long. + char type[256]; + unsigned char *value; // uint8_t* + int size; + int pad0; +} EXRAttribute; + +typedef struct _EXRChannelInfo { + char name[256]; // less than 255 bytes long + int pixel_type; + int x_sampling; + int y_sampling; + unsigned char p_linear; + unsigned char pad[3]; +} EXRChannelInfo; + +typedef struct _EXRTile { + int offset_x; + int offset_y; + int level_x; + int level_y; + + int width; // actual width in a tile. + int height; // actual height int a tile. + + unsigned char **images; // image[channels][pixels] +} EXRTile; + +typedef struct _EXRHeader { + float pixel_aspect_ratio; + int line_order; + int data_window[4]; + int display_window[4]; + float screen_window_center[2]; + float screen_window_width; + + int chunk_count; + + // Properties for tiled format(`tiledesc`). + int tiled; + int tile_size_x; + int tile_size_y; + int tile_level_mode; + int tile_rounding_mode; + + int long_name; + int non_image; + int multipart; + unsigned int header_len; + + // Custom attributes(exludes required attributes(e.g. `channels`, + // `compression`, etc) + int num_custom_attributes; + EXRAttribute *custom_attributes; // array of EXRAttribute. size = + // `num_custom_attributes`. + + EXRChannelInfo *channels; // [num_channels] + + int *pixel_types; // Loaded pixel type(TINYEXR_PIXELTYPE_*) of `images` for + // each channel. This is overwritten with `requested_pixel_types` when + // loading. + int num_channels; + + int compression_type; // compression type(TINYEXR_COMPRESSIONTYPE_*) + int *requested_pixel_types; // Filled initially by + // ParseEXRHeaderFrom(Meomory|File), then users + // can edit it(only valid for HALF pixel type + // channel) + +} EXRHeader; + +typedef struct _EXRMultiPartHeader { + int num_headers; + EXRHeader *headers; + +} EXRMultiPartHeader; + +typedef struct _EXRImage { + EXRTile *tiles; // Tiled pixel data. The application must reconstruct image + // from tiles manually. NULL if scanline format. + unsigned char **images; // image[channels][pixels]. NULL if tiled format. + + int width; + int height; + int num_channels; + + // Properties for tile format. + int num_tiles; + +} EXRImage; + +typedef struct _EXRMultiPartImage { + int num_images; + EXRImage *images; + +} EXRMultiPartImage; + +typedef struct _DeepImage { + const char **channel_names; + float ***image; // image[channels][scanlines][samples] + int **offset_table; // offset_table[scanline][offsets] + int num_channels; + int width; + int height; + int pad0; +} DeepImage; + +// @deprecated { For backward compatibility. Not recommended to use. } +// Loads single-frame OpenEXR image. Assume EXR image contains A(single channel +// alpha) or RGB(A) channels. +// Application must free image data as returned by `out_rgba` +// Result image format is: float x RGBA x width x hight +// Returns negative value and may set error string in `err` when there's an +// error +extern int LoadEXR(float **out_rgba, int *width, int *height, + const char *filename, const char **err); + +// Loads single-frame OpenEXR image by specifing layer name. Assume EXR image contains A(single channel +// alpha) or RGB(A) channels. +// Application must free image data as returned by `out_rgba` +// Result image format is: float x RGBA x width x hight +// Returns negative value and may set error string in `err` when there's an +// error +// When the specified layer name is not found in the EXR file, the function will return `TINYEXR_ERROR_LAYER_NOT_FOUND`. +extern int LoadEXRWithLayer(float **out_rgba, int *width, int *height, + const char *filename, const char *layer_name, const char **err); + +// +// Get layer infos from EXR file. +// +// @param[out] layer_names List of layer names. Application must free memory after using this. +// @param[out] num_layers The number of layers +// @param[out] err Error string(wll be filled when the function returns error code). Free it using FreeEXRErrorMessage after using this value. +// +// @return TINYEXR_SUCCEES upon success. +// +extern int EXRLayers(const char *filename, const char **layer_names[], int *num_layers, const char **err); + +// @deprecated { to be removed. } +// Simple wrapper API for ParseEXRHeaderFromFile. +// checking given file is a EXR file(by just look up header) +// @return TINYEXR_SUCCEES for EXR image, TINYEXR_ERROR_INVALID_HEADER for +// others +extern int IsEXR(const char *filename); + +// @deprecated { to be removed. } +// Saves single-frame OpenEXR image. Assume EXR image contains RGB(A) channels. +// components must be 1(Grayscale), 3(RGB) or 4(RGBA). +// Input image format is: `float x width x height`, or `float x RGB(A) x width x +// hight` +// Save image as fp16(HALF) format when `save_as_fp16` is positive non-zero +// value. +// Save image as fp32(FLOAT) format when `save_as_fp16` is 0. +// Use ZIP compression by default. +// Returns negative value and may set error string in `err` when there's an +// error +extern int SaveEXR(const float *data, const int width, const int height, + const int components, const int save_as_fp16, + const char *filename, const char **err); + +// Initialize EXRHeader struct +extern void InitEXRHeader(EXRHeader *exr_header); + +// Initialize EXRImage struct +extern void InitEXRImage(EXRImage *exr_image); + +// Free's internal data of EXRHeader struct +extern int FreeEXRHeader(EXRHeader *exr_header); + +// Free's internal data of EXRImage struct +extern int FreeEXRImage(EXRImage *exr_image); + +// Free's error message +extern void FreeEXRErrorMessage(const char *msg); + +// Parse EXR version header of a file. +extern int ParseEXRVersionFromFile(EXRVersion *version, const char *filename); + +// Parse EXR version header from memory-mapped EXR data. +extern int ParseEXRVersionFromMemory(EXRVersion *version, + const unsigned char *memory, size_t size); + +// Parse single-part OpenEXR header from a file and initialize `EXRHeader`. +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int ParseEXRHeaderFromFile(EXRHeader *header, const EXRVersion *version, + const char *filename, const char **err); + +// Parse single-part OpenEXR header from a memory and initialize `EXRHeader`. +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int ParseEXRHeaderFromMemory(EXRHeader *header, + const EXRVersion *version, + const unsigned char *memory, size_t size, + const char **err); + +// Parse multi-part OpenEXR headers from a file and initialize `EXRHeader*` +// array. +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int ParseEXRMultipartHeaderFromFile(EXRHeader ***headers, + int *num_headers, + const EXRVersion *version, + const char *filename, + const char **err); + +// Parse multi-part OpenEXR headers from a memory and initialize `EXRHeader*` +// array +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int ParseEXRMultipartHeaderFromMemory(EXRHeader ***headers, + int *num_headers, + const EXRVersion *version, + const unsigned char *memory, + size_t size, const char **err); + +// Loads single-part OpenEXR image from a file. +// Application must setup `ParseEXRHeaderFromFile` before calling this function. +// Application can free EXRImage using `FreeEXRImage` +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int LoadEXRImageFromFile(EXRImage *image, const EXRHeader *header, + const char *filename, const char **err); + +// Loads single-part OpenEXR image from a memory. +// Application must setup `EXRHeader` with +// `ParseEXRHeaderFromMemory` before calling this function. +// Application can free EXRImage using `FreeEXRImage` +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int LoadEXRImageFromMemory(EXRImage *image, const EXRHeader *header, + const unsigned char *memory, + const size_t size, const char **err); + +// Loads multi-part OpenEXR image from a file. +// Application must setup `ParseEXRMultipartHeaderFromFile` before calling this +// function. +// Application can free EXRImage using `FreeEXRImage` +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int LoadEXRMultipartImageFromFile(EXRImage *images, + const EXRHeader **headers, + unsigned int num_parts, + const char *filename, + const char **err); + +// Loads multi-part OpenEXR image from a memory. +// Application must setup `EXRHeader*` array with +// `ParseEXRMultipartHeaderFromMemory` before calling this function. +// Application can free EXRImage using `FreeEXRImage` +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int LoadEXRMultipartImageFromMemory(EXRImage *images, + const EXRHeader **headers, + unsigned int num_parts, + const unsigned char *memory, + const size_t size, const char **err); + +// Saves multi-channel, single-frame OpenEXR image to a file. +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int SaveEXRImageToFile(const EXRImage *image, + const EXRHeader *exr_header, const char *filename, + const char **err); + +// Saves multi-channel, single-frame OpenEXR image to a memory. +// Image is compressed using EXRImage.compression value. +// Return the number of bytes if success. +// Return zero and will set error string in `err` when there's an +// error. +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern size_t SaveEXRImageToMemory(const EXRImage *image, + const EXRHeader *exr_header, + unsigned char **memory, const char **err); + +// Loads single-frame OpenEXR deep image. +// Application must free memory of variables in DeepImage(image, offset_table) +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int LoadDeepEXR(DeepImage *out_image, const char *filename, + const char **err); + +// NOT YET IMPLEMENTED: +// Saves single-frame OpenEXR deep image. +// Returns negative value and may set error string in `err` when there's an +// error +// extern int SaveDeepEXR(const DeepImage *in_image, const char *filename, +// const char **err); + +// NOT YET IMPLEMENTED: +// Loads multi-part OpenEXR deep image. +// Application must free memory of variables in DeepImage(image, offset_table) +// extern int LoadMultiPartDeepEXR(DeepImage **out_image, int num_parts, const +// char *filename, +// const char **err); + +// For emscripten. +// Loads single-frame OpenEXR image from memory. Assume EXR image contains +// RGB(A) channels. +// Returns negative value and may set error string in `err` when there's an +// error +// When there was an error message, Application must free `err` with +// FreeEXRErrorMessage() +extern int LoadEXRFromMemory(float **out_rgba, int *width, int *height, + const unsigned char *memory, size_t size, + const char **err); + +#ifdef __cplusplus +} +#endif + +#endif // TINYEXR_H_ + +#ifdef TINYEXR_IMPLEMENTATION +#ifndef TINYEXR_IMPLEMENTATION_DEIFNED +#define TINYEXR_IMPLEMENTATION_DEIFNED + +#include +#include +#include +#include +#include + +// #include // debug + +#include +#include +#include + +#if __cplusplus > 199711L +// C++11 +#include + +#if TINYEXR_USE_THREAD +#include +#include +#endif + +#endif // __cplusplus > 199711L + +#if TINYEXR_USE_OPENMP +#include +#endif + +#if TINYEXR_USE_MINIZ +#else +// Issue #46. Please include your own zlib-compatible API header before +// including `tinyexr.h` +//#include "zlib.h" +#endif + +#if TINYEXR_USE_ZFP +#include "zfp.h" +#endif + +namespace tinyexr { + +#if __cplusplus > 199711L +// C++11 +typedef uint64_t tinyexr_uint64; +typedef int64_t tinyexr_int64; +#else +// Although `long long` is not a standard type pre C++11, assume it is defined +// as a compiler's extension. +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wc++11-long-long" +#endif +typedef unsigned long long tinyexr_uint64; +typedef long long tinyexr_int64; +#ifdef __clang__ +#pragma clang diagnostic pop +#endif +#endif + +#if TINYEXR_USE_MINIZ + +namespace miniz { + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wc++11-long-long" +#pragma clang diagnostic ignored "-Wold-style-cast" +#pragma clang diagnostic ignored "-Wpadded" +#pragma clang diagnostic ignored "-Wsign-conversion" +#pragma clang diagnostic ignored "-Wc++11-extensions" +#pragma clang diagnostic ignored "-Wconversion" +#pragma clang diagnostic ignored "-Wunused-function" +#pragma clang diagnostic ignored "-Wc++98-compat-pedantic" +#pragma clang diagnostic ignored "-Wundef" + +#if __has_warning("-Wcomma") +#pragma clang diagnostic ignored "-Wcomma" +#endif + +#if __has_warning("-Wmacro-redefined") +#pragma clang diagnostic ignored "-Wmacro-redefined" +#endif + +#if __has_warning("-Wcast-qual") +#pragma clang diagnostic ignored "-Wcast-qual" +#endif + +#if __has_warning("-Wzero-as-null-pointer-constant") +#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant" +#endif + +#if __has_warning("-Wtautological-constant-compare") +#pragma clang diagnostic ignored "-Wtautological-constant-compare" +#endif + +#if __has_warning("-Wextra-semi-stmt") +#pragma clang diagnostic ignored "-Wextra-semi-stmt" +#endif + +#endif + +/* miniz.c v1.15 - public domain deflate/inflate, zlib-subset, ZIP + reading/writing/appending, PNG writing + See "unlicense" statement at the end of this file. + Rich Geldreich , last updated Oct. 13, 2013 + Implements RFC 1950: http://www.ietf.org/rfc/rfc1950.txt and RFC 1951: + http://www.ietf.org/rfc/rfc1951.txt + + Most API's defined in miniz.c are optional. For example, to disable the + archive related functions just define + MINIZ_NO_ARCHIVE_APIS, or to get rid of all stdio usage define MINIZ_NO_STDIO + (see the list below for more macros). + + * Change History + 10/13/13 v1.15 r4 - Interim bugfix release while I work on the next major + release with Zip64 support (almost there!): + - Critical fix for the MZ_ZIP_FLAG_DO_NOT_SORT_CENTRAL_DIRECTORY bug + (thanks kahmyong.moon@hp.com) which could cause locate files to not find + files. This bug + would only have occured in earlier versions if you explicitly used this + flag, OR if you used mz_zip_extract_archive_file_to_heap() or + mz_zip_add_mem_to_archive_file_in_place() + (which used this flag). If you can't switch to v1.15 but want to fix + this bug, just remove the uses of this flag from both helper funcs (and of + course don't use the flag). + - Bugfix in mz_zip_reader_extract_to_mem_no_alloc() from kymoon when + pUser_read_buf is not NULL and compressed size is > uncompressed size + - Fixing mz_zip_reader_extract_*() funcs so they don't try to extract + compressed data from directory entries, to account for weird zipfiles which + contain zero-size compressed data on dir entries. + Hopefully this fix won't cause any issues on weird zip archives, + because it assumes the low 16-bits of zip external attributes are DOS + attributes (which I believe they always are in practice). + - Fixing mz_zip_reader_is_file_a_directory() so it doesn't check the + internal attributes, just the filename and external attributes + - mz_zip_reader_init_file() - missing MZ_FCLOSE() call if the seek failed + - Added cmake support for Linux builds which builds all the examples, + tested with clang v3.3 and gcc v4.6. + - Clang fix for tdefl_write_image_to_png_file_in_memory() from toffaletti + - Merged MZ_FORCEINLINE fix from hdeanclark + - Fix include before config #ifdef, thanks emil.brink + - Added tdefl_write_image_to_png_file_in_memory_ex(): supports Y flipping + (super useful for OpenGL apps), and explicit control over the compression + level (so you can + set it to 1 for real-time compression). + - Merged in some compiler fixes from paulharris's github repro. + - Retested this build under Windows (VS 2010, including static analysis), + tcc 0.9.26, gcc v4.6 and clang v3.3. + - Added example6.c, which dumps an image of the mandelbrot set to a PNG + file. + - Modified example2 to help test the + MZ_ZIP_FLAG_DO_NOT_SORT_CENTRAL_DIRECTORY flag more. + - In r3: Bugfix to mz_zip_writer_add_file() found during merge: Fix + possible src file fclose() leak if alignment bytes+local header file write + faiiled + - In r4: Minor bugfix to mz_zip_writer_add_from_zip_reader(): + Was pushing the wrong central dir header offset, appears harmless in this + release, but it became a problem in the zip64 branch + 5/20/12 v1.14 - MinGW32/64 GCC 4.6.1 compiler fixes: added MZ_FORCEINLINE, + #include (thanks fermtect). + 5/19/12 v1.13 - From jason@cornsyrup.org and kelwert@mtu.edu - Fix + mz_crc32() so it doesn't compute the wrong CRC-32's when mz_ulong is 64-bit. + - Temporarily/locally slammed in "typedef unsigned long mz_ulong" and + re-ran a randomized regression test on ~500k files. + - Eliminated a bunch of warnings when compiling with GCC 32-bit/64. + - Ran all examples, miniz.c, and tinfl.c through MSVC 2008's /analyze + (static analysis) option and fixed all warnings (except for the silly + "Use of the comma-operator in a tested expression.." analysis warning, + which I purposely use to work around a MSVC compiler warning). + - Created 32-bit and 64-bit Codeblocks projects/workspace. Built and + tested Linux executables. The codeblocks workspace is compatible with + Linux+Win32/x64. + - Added miniz_tester solution/project, which is a useful little app + derived from LZHAM's tester app that I use as part of the regression test. + - Ran miniz.c and tinfl.c through another series of regression testing on + ~500,000 files and archives. + - Modified example5.c so it purposely disables a bunch of high-level + functionality (MINIZ_NO_STDIO, etc.). (Thanks to corysama for the + MINIZ_NO_STDIO bug report.) + - Fix ftell() usage in examples so they exit with an error on files which + are too large (a limitation of the examples, not miniz itself). + 4/12/12 v1.12 - More comments, added low-level example5.c, fixed a couple + minor level_and_flags issues in the archive API's. + level_and_flags can now be set to MZ_DEFAULT_COMPRESSION. Thanks to Bruce + Dawson for the feedback/bug report. + 5/28/11 v1.11 - Added statement from unlicense.org + 5/27/11 v1.10 - Substantial compressor optimizations: + - Level 1 is now ~4x faster than before. The L1 compressor's throughput + now varies between 70-110MB/sec. on a + - Core i7 (actual throughput varies depending on the type of data, and x64 + vs. x86). + - Improved baseline L2-L9 compression perf. Also, greatly improved + compression perf. issues on some file types. + - Refactored the compression code for better readability and + maintainability. + - Added level 10 compression level (L10 has slightly better ratio than + level 9, but could have a potentially large + drop in throughput on some files). + 5/15/11 v1.09 - Initial stable release. + + * Low-level Deflate/Inflate implementation notes: + + Compression: Use the "tdefl" API's. The compressor supports raw, static, + and dynamic blocks, lazy or + greedy parsing, match length filtering, RLE-only, and Huffman-only streams. + It performs and compresses + approximately as well as zlib. + + Decompression: Use the "tinfl" API's. The entire decompressor is + implemented as a single function + coroutine: see tinfl_decompress(). It supports decompression into a 32KB + (or larger power of 2) wrapping buffer, or into a memory + block large enough to hold the entire file. + + The low-level tdefl/tinfl API's do not make any use of dynamic memory + allocation. + + * zlib-style API notes: + + miniz.c implements a fairly large subset of zlib. There's enough + functionality present for it to be a drop-in + zlib replacement in many apps: + The z_stream struct, optional memory allocation callbacks + deflateInit/deflateInit2/deflate/deflateReset/deflateEnd/deflateBound + inflateInit/inflateInit2/inflate/inflateEnd + compress, compress2, compressBound, uncompress + CRC-32, Adler-32 - Using modern, minimal code size, CPU cache friendly + routines. + Supports raw deflate streams or standard zlib streams with adler-32 + checking. + + Limitations: + The callback API's are not implemented yet. No support for gzip headers or + zlib static dictionaries. + I've tried to closely emulate zlib's various flavors of stream flushing + and return status codes, but + there are no guarantees that miniz.c pulls this off perfectly. + + * PNG writing: See the tdefl_write_image_to_png_file_in_memory() function, + originally written by + Alex Evans. Supports 1-4 bytes/pixel images. + + * ZIP archive API notes: + + The ZIP archive API's where designed with simplicity and efficiency in + mind, with just enough abstraction to + get the job done with minimal fuss. There are simple API's to retrieve file + information, read files from + existing archives, create new archives, append new files to existing + archives, or clone archive data from + one archive to another. It supports archives located in memory or the heap, + on disk (using stdio.h), + or you can specify custom file read/write callbacks. + + - Archive reading: Just call this function to read a single file from a + disk archive: + + void *mz_zip_extract_archive_file_to_heap(const char *pZip_filename, const + char *pArchive_name, + size_t *pSize, mz_uint zip_flags); + + For more complex cases, use the "mz_zip_reader" functions. Upon opening an + archive, the entire central + directory is located and read as-is into memory, and subsequent file access + only occurs when reading individual files. + + - Archives file scanning: The simple way is to use this function to scan a + loaded archive for a specific file: + + int mz_zip_reader_locate_file(mz_zip_archive *pZip, const char *pName, + const char *pComment, mz_uint flags); + + The locate operation can optionally check file comments too, which (as one + example) can be used to identify + multiple versions of the same file in an archive. This function uses a + simple linear search through the central + directory, so it's not very fast. + + Alternately, you can iterate through all the files in an archive (using + mz_zip_reader_get_num_files()) and + retrieve detailed info on each file by calling mz_zip_reader_file_stat(). + + - Archive creation: Use the "mz_zip_writer" functions. The ZIP writer + immediately writes compressed file data + to disk and builds an exact image of the central directory in memory. The + central directory image is written + all at once at the end of the archive file when the archive is finalized. + + The archive writer can optionally align each file's local header and file + data to any power of 2 alignment, + which can be useful when the archive will be read from optical media. Also, + the writer supports placing + arbitrary data blobs at the very beginning of ZIP archives. Archives + written using either feature are still + readable by any ZIP tool. + + - Archive appending: The simple way to add a single file to an archive is + to call this function: + + mz_bool mz_zip_add_mem_to_archive_file_in_place(const char *pZip_filename, + const char *pArchive_name, + const void *pBuf, size_t buf_size, const void *pComment, mz_uint16 + comment_size, mz_uint level_and_flags); + + The archive will be created if it doesn't already exist, otherwise it'll be + appended to. + Note the appending is done in-place and is not an atomic operation, so if + something goes wrong + during the operation it's possible the archive could be left without a + central directory (although the local + file headers and file data will be fine, so the archive will be + recoverable). + + For more complex archive modification scenarios: + 1. The safest way is to use a mz_zip_reader to read the existing archive, + cloning only those bits you want to + preserve into a new archive using using the + mz_zip_writer_add_from_zip_reader() function (which compiles the + compressed file data as-is). When you're done, delete the old archive and + rename the newly written archive, and + you're done. This is safe but requires a bunch of temporary disk space or + heap memory. + + 2. Or, you can convert an mz_zip_reader in-place to an mz_zip_writer using + mz_zip_writer_init_from_reader(), + append new files as needed, then finalize the archive which will write an + updated central directory to the + original archive. (This is basically what + mz_zip_add_mem_to_archive_file_in_place() does.) There's a + possibility that the archive's central directory could be lost with this + method if anything goes wrong, though. + + - ZIP archive support limitations: + No zip64 or spanning support. Extraction functions can only handle + unencrypted, stored or deflated files. + Requires streams capable of seeking. + + * This is a header file library, like stb_image.c. To get only a header file, + either cut and paste the + below header, or create miniz.h, #define MINIZ_HEADER_FILE_ONLY, and then + include miniz.c from it. + + * Important: For best perf. be sure to customize the below macros for your + target platform: + #define MINIZ_USE_UNALIGNED_LOADS_AND_STORES 1 + #define MINIZ_LITTLE_ENDIAN 1 + #define MINIZ_HAS_64BIT_REGISTERS 1 + + * On platforms using glibc, Be sure to "#define _LARGEFILE64_SOURCE 1" before + including miniz.c to ensure miniz + uses the 64-bit variants: fopen64(), stat64(), etc. Otherwise you won't be + able to process large files + (i.e. 32-bit stat() fails for me on files > 0x7FFFFFFF bytes). +*/ + +#ifndef MINIZ_HEADER_INCLUDED +#define MINIZ_HEADER_INCLUDED + +//#include + +// Defines to completely disable specific portions of miniz.c: +// If all macros here are defined the only functionality remaining will be +// CRC-32, adler-32, tinfl, and tdefl. + +// Define MINIZ_NO_STDIO to disable all usage and any functions which rely on +// stdio for file I/O. +//#define MINIZ_NO_STDIO + +// If MINIZ_NO_TIME is specified then the ZIP archive functions will not be able +// to get the current time, or +// get/set file times, and the C run-time funcs that get/set times won't be +// called. +// The current downside is the times written to your archives will be from 1979. +#define MINIZ_NO_TIME + +// Define MINIZ_NO_ARCHIVE_APIS to disable all ZIP archive API's. +#define MINIZ_NO_ARCHIVE_APIS + +// Define MINIZ_NO_ARCHIVE_APIS to disable all writing related ZIP archive +// API's. +//#define MINIZ_NO_ARCHIVE_WRITING_APIS + +// Define MINIZ_NO_ZLIB_APIS to remove all ZLIB-style compression/decompression +// API's. +//#define MINIZ_NO_ZLIB_APIS + +// Define MINIZ_NO_ZLIB_COMPATIBLE_NAME to disable zlib names, to prevent +// conflicts against stock zlib. +//#define MINIZ_NO_ZLIB_COMPATIBLE_NAMES + +// Define MINIZ_NO_MALLOC to disable all calls to malloc, free, and realloc. +// Note if MINIZ_NO_MALLOC is defined then the user must always provide custom +// user alloc/free/realloc +// callbacks to the zlib and archive API's, and a few stand-alone helper API's +// which don't provide custom user +// functions (such as tdefl_compress_mem_to_heap() and +// tinfl_decompress_mem_to_heap()) won't work. +//#define MINIZ_NO_MALLOC + +#if defined(__TINYC__) && (defined(__linux) || defined(__linux__)) +// TODO: Work around "error: include file 'sys\utime.h' when compiling with tcc +// on Linux +#define MINIZ_NO_TIME +#endif + +#if !defined(MINIZ_NO_TIME) && !defined(MINIZ_NO_ARCHIVE_APIS) +//#include +#endif + +#if defined(_M_IX86) || defined(_M_X64) || defined(__i386__) || \ + defined(__i386) || defined(__i486__) || defined(__i486) || \ + defined(i386) || defined(__ia64__) || defined(__x86_64__) +// MINIZ_X86_OR_X64_CPU is only used to help set the below macros. +#define MINIZ_X86_OR_X64_CPU 1 +#endif + +#if defined(__sparcv9) +// Big endian +#else +#if (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) || MINIZ_X86_OR_X64_CPU +// Set MINIZ_LITTLE_ENDIAN to 1 if the processor is little endian. +#define MINIZ_LITTLE_ENDIAN 1 +#endif +#endif + +#if MINIZ_X86_OR_X64_CPU +// Set MINIZ_USE_UNALIGNED_LOADS_AND_STORES to 1 on CPU's that permit efficient +// integer loads and stores from unaligned addresses. +//#define MINIZ_USE_UNALIGNED_LOADS_AND_STORES 1 +#define MINIZ_USE_UNALIGNED_LOADS_AND_STORES \ + 0 // disable to suppress compiler warnings +#endif + +#if defined(_M_X64) || defined(_WIN64) || defined(__MINGW64__) || \ + defined(_LP64) || defined(__LP64__) || defined(__ia64__) || \ + defined(__x86_64__) +// Set MINIZ_HAS_64BIT_REGISTERS to 1 if operations on 64-bit integers are +// reasonably fast (and don't involve compiler generated calls to helper +// functions). +#define MINIZ_HAS_64BIT_REGISTERS 1 +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +// ------------------- zlib-style API Definitions. + +// For more compatibility with zlib, miniz.c uses unsigned long for some +// parameters/struct members. Beware: mz_ulong can be either 32 or 64-bits! +typedef unsigned long mz_ulong; + +// mz_free() internally uses the MZ_FREE() macro (which by default calls free() +// unless you've modified the MZ_MALLOC macro) to release a block allocated from +// the heap. +void mz_free(void *p); + +#define MZ_ADLER32_INIT (1) +// mz_adler32() returns the initial adler-32 value to use when called with +// ptr==NULL. +mz_ulong mz_adler32(mz_ulong adler, const unsigned char *ptr, size_t buf_len); + +#define MZ_CRC32_INIT (0) +// mz_crc32() returns the initial CRC-32 value to use when called with +// ptr==NULL. +mz_ulong mz_crc32(mz_ulong crc, const unsigned char *ptr, size_t buf_len); + +// Compression strategies. +enum { + MZ_DEFAULT_STRATEGY = 0, + MZ_FILTERED = 1, + MZ_HUFFMAN_ONLY = 2, + MZ_RLE = 3, + MZ_FIXED = 4 +}; + +// Method +#define MZ_DEFLATED 8 + +#ifndef MINIZ_NO_ZLIB_APIS + +// Heap allocation callbacks. +// Note that mz_alloc_func parameter types purpsosely differ from zlib's: +// items/size is size_t, not unsigned long. +typedef void *(*mz_alloc_func)(void *opaque, size_t items, size_t size); +typedef void (*mz_free_func)(void *opaque, void *address); +typedef void *(*mz_realloc_func)(void *opaque, void *address, size_t items, + size_t size); + +#define MZ_VERSION "9.1.15" +#define MZ_VERNUM 0x91F0 +#define MZ_VER_MAJOR 9 +#define MZ_VER_MINOR 1 +#define MZ_VER_REVISION 15 +#define MZ_VER_SUBREVISION 0 + +// Flush values. For typical usage you only need MZ_NO_FLUSH and MZ_FINISH. The +// other values are for advanced use (refer to the zlib docs). +enum { + MZ_NO_FLUSH = 0, + MZ_PARTIAL_FLUSH = 1, + MZ_SYNC_FLUSH = 2, + MZ_FULL_FLUSH = 3, + MZ_FINISH = 4, + MZ_BLOCK = 5 +}; + +// Return status codes. MZ_PARAM_ERROR is non-standard. +enum { + MZ_OK = 0, + MZ_STREAM_END = 1, + MZ_NEED_DICT = 2, + MZ_ERRNO = -1, + MZ_STREAM_ERROR = -2, + MZ_DATA_ERROR = -3, + MZ_MEM_ERROR = -4, + MZ_BUF_ERROR = -5, + MZ_VERSION_ERROR = -6, + MZ_PARAM_ERROR = -10000 +}; + +// Compression levels: 0-9 are the standard zlib-style levels, 10 is best +// possible compression (not zlib compatible, and may be very slow), +// MZ_DEFAULT_COMPRESSION=MZ_DEFAULT_LEVEL. +enum { + MZ_NO_COMPRESSION = 0, + MZ_BEST_SPEED = 1, + MZ_BEST_COMPRESSION = 9, + MZ_UBER_COMPRESSION = 10, + MZ_DEFAULT_LEVEL = 6, + MZ_DEFAULT_COMPRESSION = -1 +}; + +// Window bits +#define MZ_DEFAULT_WINDOW_BITS 15 + +struct mz_internal_state; + +// Compression/decompression stream struct. +typedef struct mz_stream_s { + const unsigned char *next_in; // pointer to next byte to read + unsigned int avail_in; // number of bytes available at next_in + mz_ulong total_in; // total number of bytes consumed so far + + unsigned char *next_out; // pointer to next byte to write + unsigned int avail_out; // number of bytes that can be written to next_out + mz_ulong total_out; // total number of bytes produced so far + + char *msg; // error msg (unused) + struct mz_internal_state *state; // internal state, allocated by zalloc/zfree + + mz_alloc_func + zalloc; // optional heap allocation function (defaults to malloc) + mz_free_func zfree; // optional heap free function (defaults to free) + void *opaque; // heap alloc function user pointer + + int data_type; // data_type (unused) + mz_ulong adler; // adler32 of the source or uncompressed data + mz_ulong reserved; // not used +} mz_stream; + +typedef mz_stream *mz_streamp; + +// Returns the version string of miniz.c. +const char *mz_version(void); + +// mz_deflateInit() initializes a compressor with default options: +// Parameters: +// pStream must point to an initialized mz_stream struct. +// level must be between [MZ_NO_COMPRESSION, MZ_BEST_COMPRESSION]. +// level 1 enables a specially optimized compression function that's been +// optimized purely for performance, not ratio. +// (This special func. is currently only enabled when +// MINIZ_USE_UNALIGNED_LOADS_AND_STORES and MINIZ_LITTLE_ENDIAN are defined.) +// Return values: +// MZ_OK on success. +// MZ_STREAM_ERROR if the stream is bogus. +// MZ_PARAM_ERROR if the input parameters are bogus. +// MZ_MEM_ERROR on out of memory. +int mz_deflateInit(mz_streamp pStream, int level); + +// mz_deflateInit2() is like mz_deflate(), except with more control: +// Additional parameters: +// method must be MZ_DEFLATED +// window_bits must be MZ_DEFAULT_WINDOW_BITS (to wrap the deflate stream with +// zlib header/adler-32 footer) or -MZ_DEFAULT_WINDOW_BITS (raw deflate/no +// header or footer) +// mem_level must be between [1, 9] (it's checked but ignored by miniz.c) +int mz_deflateInit2(mz_streamp pStream, int level, int method, int window_bits, + int mem_level, int strategy); + +// Quickly resets a compressor without having to reallocate anything. Same as +// calling mz_deflateEnd() followed by mz_deflateInit()/mz_deflateInit2(). +int mz_deflateReset(mz_streamp pStream); + +// mz_deflate() compresses the input to output, consuming as much of the input +// and producing as much output as possible. +// Parameters: +// pStream is the stream to read from and write to. You must initialize/update +// the next_in, avail_in, next_out, and avail_out members. +// flush may be MZ_NO_FLUSH, MZ_PARTIAL_FLUSH/MZ_SYNC_FLUSH, MZ_FULL_FLUSH, or +// MZ_FINISH. +// Return values: +// MZ_OK on success (when flushing, or if more input is needed but not +// available, and/or there's more output to be written but the output buffer +// is full). +// MZ_STREAM_END if all input has been consumed and all output bytes have been +// written. Don't call mz_deflate() on the stream anymore. +// MZ_STREAM_ERROR if the stream is bogus. +// MZ_PARAM_ERROR if one of the parameters is invalid. +// MZ_BUF_ERROR if no forward progress is possible because the input and/or +// output buffers are empty. (Fill up the input buffer or free up some output +// space and try again.) +int mz_deflate(mz_streamp pStream, int flush); + +// mz_deflateEnd() deinitializes a compressor: +// Return values: +// MZ_OK on success. +// MZ_STREAM_ERROR if the stream is bogus. +int mz_deflateEnd(mz_streamp pStream); + +// mz_deflateBound() returns a (very) conservative upper bound on the amount of +// data that could be generated by deflate(), assuming flush is set to only +// MZ_NO_FLUSH or MZ_FINISH. +mz_ulong mz_deflateBound(mz_streamp pStream, mz_ulong source_len); + +// Single-call compression functions mz_compress() and mz_compress2(): +// Returns MZ_OK on success, or one of the error codes from mz_deflate() on +// failure. +int mz_compress(unsigned char *pDest, mz_ulong *pDest_len, + const unsigned char *pSource, mz_ulong source_len); +int mz_compress2(unsigned char *pDest, mz_ulong *pDest_len, + const unsigned char *pSource, mz_ulong source_len, int level); + +// mz_compressBound() returns a (very) conservative upper bound on the amount of +// data that could be generated by calling mz_compress(). +mz_ulong mz_compressBound(mz_ulong source_len); + +// Initializes a decompressor. +int mz_inflateInit(mz_streamp pStream); + +// mz_inflateInit2() is like mz_inflateInit() with an additional option that +// controls the window size and whether or not the stream has been wrapped with +// a zlib header/footer: +// window_bits must be MZ_DEFAULT_WINDOW_BITS (to parse zlib header/footer) or +// -MZ_DEFAULT_WINDOW_BITS (raw deflate). +int mz_inflateInit2(mz_streamp pStream, int window_bits); + +// Decompresses the input stream to the output, consuming only as much of the +// input as needed, and writing as much to the output as possible. +// Parameters: +// pStream is the stream to read from and write to. You must initialize/update +// the next_in, avail_in, next_out, and avail_out members. +// flush may be MZ_NO_FLUSH, MZ_SYNC_FLUSH, or MZ_FINISH. +// On the first call, if flush is MZ_FINISH it's assumed the input and output +// buffers are both sized large enough to decompress the entire stream in a +// single call (this is slightly faster). +// MZ_FINISH implies that there are no more source bytes available beside +// what's already in the input buffer, and that the output buffer is large +// enough to hold the rest of the decompressed data. +// Return values: +// MZ_OK on success. Either more input is needed but not available, and/or +// there's more output to be written but the output buffer is full. +// MZ_STREAM_END if all needed input has been consumed and all output bytes +// have been written. For zlib streams, the adler-32 of the decompressed data +// has also been verified. +// MZ_STREAM_ERROR if the stream is bogus. +// MZ_DATA_ERROR if the deflate stream is invalid. +// MZ_PARAM_ERROR if one of the parameters is invalid. +// MZ_BUF_ERROR if no forward progress is possible because the input buffer is +// empty but the inflater needs more input to continue, or if the output +// buffer is not large enough. Call mz_inflate() again +// with more input data, or with more room in the output buffer (except when +// using single call decompression, described above). +int mz_inflate(mz_streamp pStream, int flush); + +// Deinitializes a decompressor. +int mz_inflateEnd(mz_streamp pStream); + +// Single-call decompression. +// Returns MZ_OK on success, or one of the error codes from mz_inflate() on +// failure. +int mz_uncompress(unsigned char *pDest, mz_ulong *pDest_len, + const unsigned char *pSource, mz_ulong source_len); + +// Returns a string description of the specified error code, or NULL if the +// error code is invalid. +const char *mz_error(int err); + +// Redefine zlib-compatible names to miniz equivalents, so miniz.c can be used +// as a drop-in replacement for the subset of zlib that miniz.c supports. +// Define MINIZ_NO_ZLIB_COMPATIBLE_NAMES to disable zlib-compatibility if you +// use zlib in the same project. +#ifndef MINIZ_NO_ZLIB_COMPATIBLE_NAMES +typedef unsigned char Byte; +typedef unsigned int uInt; +typedef mz_ulong uLong; +typedef Byte Bytef; +typedef uInt uIntf; +typedef char charf; +typedef int intf; +typedef void *voidpf; +typedef uLong uLongf; +typedef void *voidp; +typedef void *const voidpc; +#define Z_NULL 0 +#define Z_NO_FLUSH MZ_NO_FLUSH +#define Z_PARTIAL_FLUSH MZ_PARTIAL_FLUSH +#define Z_SYNC_FLUSH MZ_SYNC_FLUSH +#define Z_FULL_FLUSH MZ_FULL_FLUSH +#define Z_FINISH MZ_FINISH +#define Z_BLOCK MZ_BLOCK +#define Z_OK MZ_OK +#define Z_STREAM_END MZ_STREAM_END +#define Z_NEED_DICT MZ_NEED_DICT +#define Z_ERRNO MZ_ERRNO +#define Z_STREAM_ERROR MZ_STREAM_ERROR +#define Z_DATA_ERROR MZ_DATA_ERROR +#define Z_MEM_ERROR MZ_MEM_ERROR +#define Z_BUF_ERROR MZ_BUF_ERROR +#define Z_VERSION_ERROR MZ_VERSION_ERROR +#define Z_PARAM_ERROR MZ_PARAM_ERROR +#define Z_NO_COMPRESSION MZ_NO_COMPRESSION +#define Z_BEST_SPEED MZ_BEST_SPEED +#define Z_BEST_COMPRESSION MZ_BEST_COMPRESSION +#define Z_DEFAULT_COMPRESSION MZ_DEFAULT_COMPRESSION +#define Z_DEFAULT_STRATEGY MZ_DEFAULT_STRATEGY +#define Z_FILTERED MZ_FILTERED +#define Z_HUFFMAN_ONLY MZ_HUFFMAN_ONLY +#define Z_RLE MZ_RLE +#define Z_FIXED MZ_FIXED +#define Z_DEFLATED MZ_DEFLATED +#define Z_DEFAULT_WINDOW_BITS MZ_DEFAULT_WINDOW_BITS +#define alloc_func mz_alloc_func +#define free_func mz_free_func +#define internal_state mz_internal_state +#define z_stream mz_stream +#define deflateInit mz_deflateInit +#define deflateInit2 mz_deflateInit2 +#define deflateReset mz_deflateReset +#define deflate mz_deflate +#define deflateEnd mz_deflateEnd +#define deflateBound mz_deflateBound +#define compress mz_compress +#define compress2 mz_compress2 +#define compressBound mz_compressBound +#define inflateInit mz_inflateInit +#define inflateInit2 mz_inflateInit2 +#define inflate mz_inflate +#define inflateEnd mz_inflateEnd +#define uncompress mz_uncompress +#define crc32 mz_crc32 +#define adler32 mz_adler32 +#define MAX_WBITS 15 +#define MAX_MEM_LEVEL 9 +#define zError mz_error +#define ZLIB_VERSION MZ_VERSION +#define ZLIB_VERNUM MZ_VERNUM +#define ZLIB_VER_MAJOR MZ_VER_MAJOR +#define ZLIB_VER_MINOR MZ_VER_MINOR +#define ZLIB_VER_REVISION MZ_VER_REVISION +#define ZLIB_VER_SUBREVISION MZ_VER_SUBREVISION +#define zlibVersion mz_version +#define zlib_version mz_version() +#endif // #ifndef MINIZ_NO_ZLIB_COMPATIBLE_NAMES + +#endif // MINIZ_NO_ZLIB_APIS + +// ------------------- Types and macros + +typedef unsigned char mz_uint8; +typedef signed short mz_int16; +typedef unsigned short mz_uint16; +typedef unsigned int mz_uint32; +typedef unsigned int mz_uint; +typedef long long mz_int64; +typedef unsigned long long mz_uint64; +typedef int mz_bool; + +#define MZ_FALSE (0) +#define MZ_TRUE (1) + +// An attempt to work around MSVC's spammy "warning C4127: conditional +// expression is constant" message. +#ifdef _MSC_VER +#define MZ_MACRO_END while (0, 0) +#else +#define MZ_MACRO_END while (0) +#endif + +// ------------------- ZIP archive reading/writing + +#ifndef MINIZ_NO_ARCHIVE_APIS + +enum { + MZ_ZIP_MAX_IO_BUF_SIZE = 64 * 1024, + MZ_ZIP_MAX_ARCHIVE_FILENAME_SIZE = 260, + MZ_ZIP_MAX_ARCHIVE_FILE_COMMENT_SIZE = 256 +}; + +typedef struct { + mz_uint32 m_file_index; + mz_uint32 m_central_dir_ofs; + mz_uint16 m_version_made_by; + mz_uint16 m_version_needed; + mz_uint16 m_bit_flag; + mz_uint16 m_method; +#ifndef MINIZ_NO_TIME + time_t m_time; +#endif + mz_uint32 m_crc32; + mz_uint64 m_comp_size; + mz_uint64 m_uncomp_size; + mz_uint16 m_internal_attr; + mz_uint32 m_external_attr; + mz_uint64 m_local_header_ofs; + mz_uint32 m_comment_size; + char m_filename[MZ_ZIP_MAX_ARCHIVE_FILENAME_SIZE]; + char m_comment[MZ_ZIP_MAX_ARCHIVE_FILE_COMMENT_SIZE]; +} mz_zip_archive_file_stat; + +typedef size_t (*mz_file_read_func)(void *pOpaque, mz_uint64 file_ofs, + void *pBuf, size_t n); +typedef size_t (*mz_file_write_func)(void *pOpaque, mz_uint64 file_ofs, + const void *pBuf, size_t n); + +struct mz_zip_internal_state_tag; +typedef struct mz_zip_internal_state_tag mz_zip_internal_state; + +typedef enum { + MZ_ZIP_MODE_INVALID = 0, + MZ_ZIP_MODE_READING = 1, + MZ_ZIP_MODE_WRITING = 2, + MZ_ZIP_MODE_WRITING_HAS_BEEN_FINALIZED = 3 +} mz_zip_mode; + +typedef struct mz_zip_archive_tag { + mz_uint64 m_archive_size; + mz_uint64 m_central_directory_file_ofs; + mz_uint m_total_files; + mz_zip_mode m_zip_mode; + + mz_uint m_file_offset_alignment; + + mz_alloc_func m_pAlloc; + mz_free_func m_pFree; + mz_realloc_func m_pRealloc; + void *m_pAlloc_opaque; + + mz_file_read_func m_pRead; + mz_file_write_func m_pWrite; + void *m_pIO_opaque; + + mz_zip_internal_state *m_pState; + +} mz_zip_archive; + +typedef enum { + MZ_ZIP_FLAG_CASE_SENSITIVE = 0x0100, + MZ_ZIP_FLAG_IGNORE_PATH = 0x0200, + MZ_ZIP_FLAG_COMPRESSED_DATA = 0x0400, + MZ_ZIP_FLAG_DO_NOT_SORT_CENTRAL_DIRECTORY = 0x0800 +} mz_zip_flags; + +// ZIP archive reading + +// Inits a ZIP archive reader. +// These functions read and validate the archive's central directory. +mz_bool mz_zip_reader_init(mz_zip_archive *pZip, mz_uint64 size, + mz_uint32 flags); +mz_bool mz_zip_reader_init_mem(mz_zip_archive *pZip, const void *pMem, + size_t size, mz_uint32 flags); + +#ifndef MINIZ_NO_STDIO +mz_bool mz_zip_reader_init_file(mz_zip_archive *pZip, const char *pFilename, + mz_uint32 flags); +#endif + +// Returns the total number of files in the archive. +mz_uint mz_zip_reader_get_num_files(mz_zip_archive *pZip); + +// Returns detailed information about an archive file entry. +mz_bool mz_zip_reader_file_stat(mz_zip_archive *pZip, mz_uint file_index, + mz_zip_archive_file_stat *pStat); + +// Determines if an archive file entry is a directory entry. +mz_bool mz_zip_reader_is_file_a_directory(mz_zip_archive *pZip, + mz_uint file_index); +mz_bool mz_zip_reader_is_file_encrypted(mz_zip_archive *pZip, + mz_uint file_index); + +// Retrieves the filename of an archive file entry. +// Returns the number of bytes written to pFilename, or if filename_buf_size is +// 0 this function returns the number of bytes needed to fully store the +// filename. +mz_uint mz_zip_reader_get_filename(mz_zip_archive *pZip, mz_uint file_index, + char *pFilename, mz_uint filename_buf_size); + +// Attempts to locates a file in the archive's central directory. +// Valid flags: MZ_ZIP_FLAG_CASE_SENSITIVE, MZ_ZIP_FLAG_IGNORE_PATH +// Returns -1 if the file cannot be found. +int mz_zip_reader_locate_file(mz_zip_archive *pZip, const char *pName, + const char *pComment, mz_uint flags); + +// Extracts a archive file to a memory buffer using no memory allocation. +mz_bool mz_zip_reader_extract_to_mem_no_alloc(mz_zip_archive *pZip, + mz_uint file_index, void *pBuf, + size_t buf_size, mz_uint flags, + void *pUser_read_buf, + size_t user_read_buf_size); +mz_bool mz_zip_reader_extract_file_to_mem_no_alloc( + mz_zip_archive *pZip, const char *pFilename, void *pBuf, size_t buf_size, + mz_uint flags, void *pUser_read_buf, size_t user_read_buf_size); + +// Extracts a archive file to a memory buffer. +mz_bool mz_zip_reader_extract_to_mem(mz_zip_archive *pZip, mz_uint file_index, + void *pBuf, size_t buf_size, + mz_uint flags); +mz_bool mz_zip_reader_extract_file_to_mem(mz_zip_archive *pZip, + const char *pFilename, void *pBuf, + size_t buf_size, mz_uint flags); + +// Extracts a archive file to a dynamically allocated heap buffer. +void *mz_zip_reader_extract_to_heap(mz_zip_archive *pZip, mz_uint file_index, + size_t *pSize, mz_uint flags); +void *mz_zip_reader_extract_file_to_heap(mz_zip_archive *pZip, + const char *pFilename, size_t *pSize, + mz_uint flags); + +// Extracts a archive file using a callback function to output the file's data. +mz_bool mz_zip_reader_extract_to_callback(mz_zip_archive *pZip, + mz_uint file_index, + mz_file_write_func pCallback, + void *pOpaque, mz_uint flags); +mz_bool mz_zip_reader_extract_file_to_callback(mz_zip_archive *pZip, + const char *pFilename, + mz_file_write_func pCallback, + void *pOpaque, mz_uint flags); + +#ifndef MINIZ_NO_STDIO +// Extracts a archive file to a disk file and sets its last accessed and +// modified times. +// This function only extracts files, not archive directory records. +mz_bool mz_zip_reader_extract_to_file(mz_zip_archive *pZip, mz_uint file_index, + const char *pDst_filename, mz_uint flags); +mz_bool mz_zip_reader_extract_file_to_file(mz_zip_archive *pZip, + const char *pArchive_filename, + const char *pDst_filename, + mz_uint flags); +#endif + +// Ends archive reading, freeing all allocations, and closing the input archive +// file if mz_zip_reader_init_file() was used. +mz_bool mz_zip_reader_end(mz_zip_archive *pZip); + +// ZIP archive writing + +#ifndef MINIZ_NO_ARCHIVE_WRITING_APIS + +// Inits a ZIP archive writer. +mz_bool mz_zip_writer_init(mz_zip_archive *pZip, mz_uint64 existing_size); +mz_bool mz_zip_writer_init_heap(mz_zip_archive *pZip, + size_t size_to_reserve_at_beginning, + size_t initial_allocation_size); + +#ifndef MINIZ_NO_STDIO +mz_bool mz_zip_writer_init_file(mz_zip_archive *pZip, const char *pFilename, + mz_uint64 size_to_reserve_at_beginning); +#endif + +// Converts a ZIP archive reader object into a writer object, to allow efficient +// in-place file appends to occur on an existing archive. +// For archives opened using mz_zip_reader_init_file, pFilename must be the +// archive's filename so it can be reopened for writing. If the file can't be +// reopened, mz_zip_reader_end() will be called. +// For archives opened using mz_zip_reader_init_mem, the memory block must be +// growable using the realloc callback (which defaults to realloc unless you've +// overridden it). +// Finally, for archives opened using mz_zip_reader_init, the mz_zip_archive's +// user provided m_pWrite function cannot be NULL. +// Note: In-place archive modification is not recommended unless you know what +// you're doing, because if execution stops or something goes wrong before +// the archive is finalized the file's central directory will be hosed. +mz_bool mz_zip_writer_init_from_reader(mz_zip_archive *pZip, + const char *pFilename); + +// Adds the contents of a memory buffer to an archive. These functions record +// the current local time into the archive. +// To add a directory entry, call this method with an archive name ending in a +// forwardslash with empty buffer. +// level_and_flags - compression level (0-10, see MZ_BEST_SPEED, +// MZ_BEST_COMPRESSION, etc.) logically OR'd with zero or more mz_zip_flags, or +// just set to MZ_DEFAULT_COMPRESSION. +mz_bool mz_zip_writer_add_mem(mz_zip_archive *pZip, const char *pArchive_name, + const void *pBuf, size_t buf_size, + mz_uint level_and_flags); +mz_bool mz_zip_writer_add_mem_ex(mz_zip_archive *pZip, + const char *pArchive_name, const void *pBuf, + size_t buf_size, const void *pComment, + mz_uint16 comment_size, + mz_uint level_and_flags, mz_uint64 uncomp_size, + mz_uint32 uncomp_crc32); + +#ifndef MINIZ_NO_STDIO +// Adds the contents of a disk file to an archive. This function also records +// the disk file's modified time into the archive. +// level_and_flags - compression level (0-10, see MZ_BEST_SPEED, +// MZ_BEST_COMPRESSION, etc.) logically OR'd with zero or more mz_zip_flags, or +// just set to MZ_DEFAULT_COMPRESSION. +mz_bool mz_zip_writer_add_file(mz_zip_archive *pZip, const char *pArchive_name, + const char *pSrc_filename, const void *pComment, + mz_uint16 comment_size, mz_uint level_and_flags); +#endif + +// Adds a file to an archive by fully cloning the data from another archive. +// This function fully clones the source file's compressed data (no +// recompression), along with its full filename, extra data, and comment fields. +mz_bool mz_zip_writer_add_from_zip_reader(mz_zip_archive *pZip, + mz_zip_archive *pSource_zip, + mz_uint file_index); + +// Finalizes the archive by writing the central directory records followed by +// the end of central directory record. +// After an archive is finalized, the only valid call on the mz_zip_archive +// struct is mz_zip_writer_end(). +// An archive must be manually finalized by calling this function for it to be +// valid. +mz_bool mz_zip_writer_finalize_archive(mz_zip_archive *pZip); +mz_bool mz_zip_writer_finalize_heap_archive(mz_zip_archive *pZip, void **pBuf, + size_t *pSize); + +// Ends archive writing, freeing all allocations, and closing the output file if +// mz_zip_writer_init_file() was used. +// Note for the archive to be valid, it must have been finalized before ending. +mz_bool mz_zip_writer_end(mz_zip_archive *pZip); + +// Misc. high-level helper functions: + +// mz_zip_add_mem_to_archive_file_in_place() efficiently (but not atomically) +// appends a memory blob to a ZIP archive. +// level_and_flags - compression level (0-10, see MZ_BEST_SPEED, +// MZ_BEST_COMPRESSION, etc.) logically OR'd with zero or more mz_zip_flags, or +// just set to MZ_DEFAULT_COMPRESSION. +mz_bool mz_zip_add_mem_to_archive_file_in_place( + const char *pZip_filename, const char *pArchive_name, const void *pBuf, + size_t buf_size, const void *pComment, mz_uint16 comment_size, + mz_uint level_and_flags); + +// Reads a single file from an archive into a heap block. +// Returns NULL on failure. +void *mz_zip_extract_archive_file_to_heap(const char *pZip_filename, + const char *pArchive_name, + size_t *pSize, mz_uint zip_flags); + +#endif // #ifndef MINIZ_NO_ARCHIVE_WRITING_APIS + +#endif // #ifndef MINIZ_NO_ARCHIVE_APIS + +// ------------------- Low-level Decompression API Definitions + +// Decompression flags used by tinfl_decompress(). +// TINFL_FLAG_PARSE_ZLIB_HEADER: If set, the input has a valid zlib header and +// ends with an adler32 checksum (it's a valid zlib stream). Otherwise, the +// input is a raw deflate stream. +// TINFL_FLAG_HAS_MORE_INPUT: If set, there are more input bytes available +// beyond the end of the supplied input buffer. If clear, the input buffer +// contains all remaining input. +// TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF: If set, the output buffer is large +// enough to hold the entire decompressed stream. If clear, the output buffer is +// at least the size of the dictionary (typically 32KB). +// TINFL_FLAG_COMPUTE_ADLER32: Force adler-32 checksum computation of the +// decompressed bytes. +enum { + TINFL_FLAG_PARSE_ZLIB_HEADER = 1, + TINFL_FLAG_HAS_MORE_INPUT = 2, + TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF = 4, + TINFL_FLAG_COMPUTE_ADLER32 = 8 +}; + +// High level decompression functions: +// tinfl_decompress_mem_to_heap() decompresses a block in memory to a heap block +// allocated via malloc(). +// On entry: +// pSrc_buf, src_buf_len: Pointer and size of the Deflate or zlib source data +// to decompress. +// On return: +// Function returns a pointer to the decompressed data, or NULL on failure. +// *pOut_len will be set to the decompressed data's size, which could be larger +// than src_buf_len on uncompressible data. +// The caller must call mz_free() on the returned block when it's no longer +// needed. +void *tinfl_decompress_mem_to_heap(const void *pSrc_buf, size_t src_buf_len, + size_t *pOut_len, int flags); + +// tinfl_decompress_mem_to_mem() decompresses a block in memory to another block +// in memory. +// Returns TINFL_DECOMPRESS_MEM_TO_MEM_FAILED on failure, or the number of bytes +// written on success. +#define TINFL_DECOMPRESS_MEM_TO_MEM_FAILED ((size_t)(-1)) +size_t tinfl_decompress_mem_to_mem(void *pOut_buf, size_t out_buf_len, + const void *pSrc_buf, size_t src_buf_len, + int flags); + +// tinfl_decompress_mem_to_callback() decompresses a block in memory to an +// internal 32KB buffer, and a user provided callback function will be called to +// flush the buffer. +// Returns 1 on success or 0 on failure. +typedef int (*tinfl_put_buf_func_ptr)(const void *pBuf, int len, void *pUser); +int tinfl_decompress_mem_to_callback(const void *pIn_buf, size_t *pIn_buf_size, + tinfl_put_buf_func_ptr pPut_buf_func, + void *pPut_buf_user, int flags); + +struct tinfl_decompressor_tag; +typedef struct tinfl_decompressor_tag tinfl_decompressor; + +// Max size of LZ dictionary. +#define TINFL_LZ_DICT_SIZE 32768 + +// Return status. +typedef enum { + TINFL_STATUS_BAD_PARAM = -3, + TINFL_STATUS_ADLER32_MISMATCH = -2, + TINFL_STATUS_FAILED = -1, + TINFL_STATUS_DONE = 0, + TINFL_STATUS_NEEDS_MORE_INPUT = 1, + TINFL_STATUS_HAS_MORE_OUTPUT = 2 +} tinfl_status; + +// Initializes the decompressor to its initial state. +#define tinfl_init(r) \ + do { \ + (r)->m_state = 0; \ + } \ + MZ_MACRO_END +#define tinfl_get_adler32(r) (r)->m_check_adler32 + +// Main low-level decompressor coroutine function. This is the only function +// actually needed for decompression. All the other functions are just +// high-level helpers for improved usability. +// This is a universal API, i.e. it can be used as a building block to build any +// desired higher level decompression API. In the limit case, it can be called +// once per every byte input or output. +tinfl_status tinfl_decompress(tinfl_decompressor *r, + const mz_uint8 *pIn_buf_next, + size_t *pIn_buf_size, mz_uint8 *pOut_buf_start, + mz_uint8 *pOut_buf_next, size_t *pOut_buf_size, + const mz_uint32 decomp_flags); + +// Internal/private bits follow. +enum { + TINFL_MAX_HUFF_TABLES = 3, + TINFL_MAX_HUFF_SYMBOLS_0 = 288, + TINFL_MAX_HUFF_SYMBOLS_1 = 32, + TINFL_MAX_HUFF_SYMBOLS_2 = 19, + TINFL_FAST_LOOKUP_BITS = 10, + TINFL_FAST_LOOKUP_SIZE = 1 << TINFL_FAST_LOOKUP_BITS +}; + +typedef struct { + mz_uint8 m_code_size[TINFL_MAX_HUFF_SYMBOLS_0]; + mz_int16 m_look_up[TINFL_FAST_LOOKUP_SIZE], + m_tree[TINFL_MAX_HUFF_SYMBOLS_0 * 2]; +} tinfl_huff_table; + +#if MINIZ_HAS_64BIT_REGISTERS +#define TINFL_USE_64BIT_BITBUF 1 +#endif + +#if TINFL_USE_64BIT_BITBUF +typedef mz_uint64 tinfl_bit_buf_t; +#define TINFL_BITBUF_SIZE (64) +#else +typedef mz_uint32 tinfl_bit_buf_t; +#define TINFL_BITBUF_SIZE (32) +#endif + +struct tinfl_decompressor_tag { + mz_uint32 m_state, m_num_bits, m_zhdr0, m_zhdr1, m_z_adler32, m_final, m_type, + m_check_adler32, m_dist, m_counter, m_num_extra, + m_table_sizes[TINFL_MAX_HUFF_TABLES]; + tinfl_bit_buf_t m_bit_buf; + size_t m_dist_from_out_buf_start; + tinfl_huff_table m_tables[TINFL_MAX_HUFF_TABLES]; + mz_uint8 m_raw_header[4], + m_len_codes[TINFL_MAX_HUFF_SYMBOLS_0 + TINFL_MAX_HUFF_SYMBOLS_1 + 137]; +}; + +// ------------------- Low-level Compression API Definitions + +// Set TDEFL_LESS_MEMORY to 1 to use less memory (compression will be slightly +// slower, and raw/dynamic blocks will be output more frequently). +#define TDEFL_LESS_MEMORY 0 + +// tdefl_init() compression flags logically OR'd together (low 12 bits contain +// the max. number of probes per dictionary search): +// TDEFL_DEFAULT_MAX_PROBES: The compressor defaults to 128 dictionary probes +// per dictionary search. 0=Huffman only, 1=Huffman+LZ (fastest/crap +// compression), 4095=Huffman+LZ (slowest/best compression). +enum { + TDEFL_HUFFMAN_ONLY = 0, + TDEFL_DEFAULT_MAX_PROBES = 128, + TDEFL_MAX_PROBES_MASK = 0xFFF +}; + +// TDEFL_WRITE_ZLIB_HEADER: If set, the compressor outputs a zlib header before +// the deflate data, and the Adler-32 of the source data at the end. Otherwise, +// you'll get raw deflate data. +// TDEFL_COMPUTE_ADLER32: Always compute the adler-32 of the input data (even +// when not writing zlib headers). +// TDEFL_GREEDY_PARSING_FLAG: Set to use faster greedy parsing, instead of more +// efficient lazy parsing. +// TDEFL_NONDETERMINISTIC_PARSING_FLAG: Enable to decrease the compressor's +// initialization time to the minimum, but the output may vary from run to run +// given the same input (depending on the contents of memory). +// TDEFL_RLE_MATCHES: Only look for RLE matches (matches with a distance of 1) +// TDEFL_FILTER_MATCHES: Discards matches <= 5 chars if enabled. +// TDEFL_FORCE_ALL_STATIC_BLOCKS: Disable usage of optimized Huffman tables. +// TDEFL_FORCE_ALL_RAW_BLOCKS: Only use raw (uncompressed) deflate blocks. +// The low 12 bits are reserved to control the max # of hash probes per +// dictionary lookup (see TDEFL_MAX_PROBES_MASK). +enum { + TDEFL_WRITE_ZLIB_HEADER = 0x01000, + TDEFL_COMPUTE_ADLER32 = 0x02000, + TDEFL_GREEDY_PARSING_FLAG = 0x04000, + TDEFL_NONDETERMINISTIC_PARSING_FLAG = 0x08000, + TDEFL_RLE_MATCHES = 0x10000, + TDEFL_FILTER_MATCHES = 0x20000, + TDEFL_FORCE_ALL_STATIC_BLOCKS = 0x40000, + TDEFL_FORCE_ALL_RAW_BLOCKS = 0x80000 +}; + +// High level compression functions: +// tdefl_compress_mem_to_heap() compresses a block in memory to a heap block +// allocated via malloc(). +// On entry: +// pSrc_buf, src_buf_len: Pointer and size of source block to compress. +// flags: The max match finder probes (default is 128) logically OR'd against +// the above flags. Higher probes are slower but improve compression. +// On return: +// Function returns a pointer to the compressed data, or NULL on failure. +// *pOut_len will be set to the compressed data's size, which could be larger +// than src_buf_len on uncompressible data. +// The caller must free() the returned block when it's no longer needed. +void *tdefl_compress_mem_to_heap(const void *pSrc_buf, size_t src_buf_len, + size_t *pOut_len, int flags); + +// tdefl_compress_mem_to_mem() compresses a block in memory to another block in +// memory. +// Returns 0 on failure. +size_t tdefl_compress_mem_to_mem(void *pOut_buf, size_t out_buf_len, + const void *pSrc_buf, size_t src_buf_len, + int flags); + +// Compresses an image to a compressed PNG file in memory. +// On entry: +// pImage, w, h, and num_chans describe the image to compress. num_chans may be +// 1, 2, 3, or 4. +// The image pitch in bytes per scanline will be w*num_chans. The leftmost +// pixel on the top scanline is stored first in memory. +// level may range from [0,10], use MZ_NO_COMPRESSION, MZ_BEST_SPEED, +// MZ_BEST_COMPRESSION, etc. or a decent default is MZ_DEFAULT_LEVEL +// If flip is true, the image will be flipped on the Y axis (useful for OpenGL +// apps). +// On return: +// Function returns a pointer to the compressed data, or NULL on failure. +// *pLen_out will be set to the size of the PNG image file. +// The caller must mz_free() the returned heap block (which will typically be +// larger than *pLen_out) when it's no longer needed. +void *tdefl_write_image_to_png_file_in_memory_ex(const void *pImage, int w, + int h, int num_chans, + size_t *pLen_out, + mz_uint level, mz_bool flip); +void *tdefl_write_image_to_png_file_in_memory(const void *pImage, int w, int h, + int num_chans, size_t *pLen_out); + +// Output stream interface. The compressor uses this interface to write +// compressed data. It'll typically be called TDEFL_OUT_BUF_SIZE at a time. +typedef mz_bool (*tdefl_put_buf_func_ptr)(const void *pBuf, int len, + void *pUser); + +// tdefl_compress_mem_to_output() compresses a block to an output stream. The +// above helpers use this function internally. +mz_bool tdefl_compress_mem_to_output(const void *pBuf, size_t buf_len, + tdefl_put_buf_func_ptr pPut_buf_func, + void *pPut_buf_user, int flags); + +enum { + TDEFL_MAX_HUFF_TABLES = 3, + TDEFL_MAX_HUFF_SYMBOLS_0 = 288, + TDEFL_MAX_HUFF_SYMBOLS_1 = 32, + TDEFL_MAX_HUFF_SYMBOLS_2 = 19, + TDEFL_LZ_DICT_SIZE = 32768, + TDEFL_LZ_DICT_SIZE_MASK = TDEFL_LZ_DICT_SIZE - 1, + TDEFL_MIN_MATCH_LEN = 3, + TDEFL_MAX_MATCH_LEN = 258 +}; + +// TDEFL_OUT_BUF_SIZE MUST be large enough to hold a single entire compressed +// output block (using static/fixed Huffman codes). +#if TDEFL_LESS_MEMORY +enum { + TDEFL_LZ_CODE_BUF_SIZE = 24 * 1024, + TDEFL_OUT_BUF_SIZE = (TDEFL_LZ_CODE_BUF_SIZE * 13) / 10, + TDEFL_MAX_HUFF_SYMBOLS = 288, + TDEFL_LZ_HASH_BITS = 12, + TDEFL_LEVEL1_HASH_SIZE_MASK = 4095, + TDEFL_LZ_HASH_SHIFT = (TDEFL_LZ_HASH_BITS + 2) / 3, + TDEFL_LZ_HASH_SIZE = 1 << TDEFL_LZ_HASH_BITS +}; +#else +enum { + TDEFL_LZ_CODE_BUF_SIZE = 64 * 1024, + TDEFL_OUT_BUF_SIZE = (TDEFL_LZ_CODE_BUF_SIZE * 13) / 10, + TDEFL_MAX_HUFF_SYMBOLS = 288, + TDEFL_LZ_HASH_BITS = 15, + TDEFL_LEVEL1_HASH_SIZE_MASK = 4095, + TDEFL_LZ_HASH_SHIFT = (TDEFL_LZ_HASH_BITS + 2) / 3, + TDEFL_LZ_HASH_SIZE = 1 << TDEFL_LZ_HASH_BITS +}; +#endif + +// The low-level tdefl functions below may be used directly if the above helper +// functions aren't flexible enough. The low-level functions don't make any heap +// allocations, unlike the above helper functions. +typedef enum { + TDEFL_STATUS_BAD_PARAM = -2, + TDEFL_STATUS_PUT_BUF_FAILED = -1, + TDEFL_STATUS_OKAY = 0, + TDEFL_STATUS_DONE = 1 +} tdefl_status; + +// Must map to MZ_NO_FLUSH, MZ_SYNC_FLUSH, etc. enums +typedef enum { + TDEFL_NO_FLUSH = 0, + TDEFL_SYNC_FLUSH = 2, + TDEFL_FULL_FLUSH = 3, + TDEFL_FINISH = 4 +} tdefl_flush; + +// tdefl's compression state structure. +typedef struct { + tdefl_put_buf_func_ptr m_pPut_buf_func; + void *m_pPut_buf_user; + mz_uint m_flags, m_max_probes[2]; + int m_greedy_parsing; + mz_uint m_adler32, m_lookahead_pos, m_lookahead_size, m_dict_size; + mz_uint8 *m_pLZ_code_buf, *m_pLZ_flags, *m_pOutput_buf, *m_pOutput_buf_end; + mz_uint m_num_flags_left, m_total_lz_bytes, m_lz_code_buf_dict_pos, m_bits_in, + m_bit_buffer; + mz_uint m_saved_match_dist, m_saved_match_len, m_saved_lit, + m_output_flush_ofs, m_output_flush_remaining, m_finished, m_block_index, + m_wants_to_finish; + tdefl_status m_prev_return_status; + const void *m_pIn_buf; + void *m_pOut_buf; + size_t *m_pIn_buf_size, *m_pOut_buf_size; + tdefl_flush m_flush; + const mz_uint8 *m_pSrc; + size_t m_src_buf_left, m_out_buf_ofs; + mz_uint8 m_dict[TDEFL_LZ_DICT_SIZE + TDEFL_MAX_MATCH_LEN - 1]; + mz_uint16 m_huff_count[TDEFL_MAX_HUFF_TABLES][TDEFL_MAX_HUFF_SYMBOLS]; + mz_uint16 m_huff_codes[TDEFL_MAX_HUFF_TABLES][TDEFL_MAX_HUFF_SYMBOLS]; + mz_uint8 m_huff_code_sizes[TDEFL_MAX_HUFF_TABLES][TDEFL_MAX_HUFF_SYMBOLS]; + mz_uint8 m_lz_code_buf[TDEFL_LZ_CODE_BUF_SIZE]; + mz_uint16 m_next[TDEFL_LZ_DICT_SIZE]; + mz_uint16 m_hash[TDEFL_LZ_HASH_SIZE]; + mz_uint8 m_output_buf[TDEFL_OUT_BUF_SIZE]; +} tdefl_compressor; + +// Initializes the compressor. +// There is no corresponding deinit() function because the tdefl API's do not +// dynamically allocate memory. +// pBut_buf_func: If NULL, output data will be supplied to the specified +// callback. In this case, the user should call the tdefl_compress_buffer() API +// for compression. +// If pBut_buf_func is NULL the user should always call the tdefl_compress() +// API. +// flags: See the above enums (TDEFL_HUFFMAN_ONLY, TDEFL_WRITE_ZLIB_HEADER, +// etc.) +tdefl_status tdefl_init(tdefl_compressor *d, + tdefl_put_buf_func_ptr pPut_buf_func, + void *pPut_buf_user, int flags); + +// Compresses a block of data, consuming as much of the specified input buffer +// as possible, and writing as much compressed data to the specified output +// buffer as possible. +tdefl_status tdefl_compress(tdefl_compressor *d, const void *pIn_buf, + size_t *pIn_buf_size, void *pOut_buf, + size_t *pOut_buf_size, tdefl_flush flush); + +// tdefl_compress_buffer() is only usable when the tdefl_init() is called with a +// non-NULL tdefl_put_buf_func_ptr. +// tdefl_compress_buffer() always consumes the entire input buffer. +tdefl_status tdefl_compress_buffer(tdefl_compressor *d, const void *pIn_buf, + size_t in_buf_size, tdefl_flush flush); + +tdefl_status tdefl_get_prev_return_status(tdefl_compressor *d); +mz_uint32 tdefl_get_adler32(tdefl_compressor *d); + +// Can't use tdefl_create_comp_flags_from_zip_params if MINIZ_NO_ZLIB_APIS isn't +// defined, because it uses some of its macros. +#ifndef MINIZ_NO_ZLIB_APIS +// Create tdefl_compress() flags given zlib-style compression parameters. +// level may range from [0,10] (where 10 is absolute max compression, but may be +// much slower on some files) +// window_bits may be -15 (raw deflate) or 15 (zlib) +// strategy may be either MZ_DEFAULT_STRATEGY, MZ_FILTERED, MZ_HUFFMAN_ONLY, +// MZ_RLE, or MZ_FIXED +mz_uint tdefl_create_comp_flags_from_zip_params(int level, int window_bits, + int strategy); +#endif // #ifndef MINIZ_NO_ZLIB_APIS + +#ifdef __cplusplus +} +#endif + +#endif // MINIZ_HEADER_INCLUDED + +// ------------------- End of Header: Implementation follows. (If you only want +// the header, define MINIZ_HEADER_FILE_ONLY.) + +#ifndef MINIZ_HEADER_FILE_ONLY + +typedef unsigned char mz_validate_uint16[sizeof(mz_uint16) == 2 ? 1 : -1]; +typedef unsigned char mz_validate_uint32[sizeof(mz_uint32) == 4 ? 1 : -1]; +typedef unsigned char mz_validate_uint64[sizeof(mz_uint64) == 8 ? 1 : -1]; + +#define MZ_ASSERT(x) TEXR_ASSERT(x) + +#ifdef MINIZ_NO_MALLOC +#define MZ_MALLOC(x) NULL +#define MZ_FREE(x) (void)x, ((void)0) +#define MZ_REALLOC(p, x) NULL +#else +#define MZ_MALLOC(x) malloc(x) +#define MZ_FREE(x) free(x) +#define MZ_REALLOC(p, x) realloc(p, x) +#endif + +#define MZ_MAX(a, b) (((a) > (b)) ? (a) : (b)) +#define MZ_MIN(a, b) (((a) < (b)) ? (a) : (b)) +#define MZ_CLEAR_OBJ(obj) memset(&(obj), 0, sizeof(obj)) + +#if MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN +#define MZ_READ_LE16(p) *((const mz_uint16 *)(p)) +#define MZ_READ_LE32(p) *((const mz_uint32 *)(p)) +#else +#define MZ_READ_LE16(p) \ + ((mz_uint32)(((const mz_uint8 *)(p))[0]) | \ + ((mz_uint32)(((const mz_uint8 *)(p))[1]) << 8U)) +#define MZ_READ_LE32(p) \ + ((mz_uint32)(((const mz_uint8 *)(p))[0]) | \ + ((mz_uint32)(((const mz_uint8 *)(p))[1]) << 8U) | \ + ((mz_uint32)(((const mz_uint8 *)(p))[2]) << 16U) | \ + ((mz_uint32)(((const mz_uint8 *)(p))[3]) << 24U)) +#endif + +#ifdef _MSC_VER +#define MZ_FORCEINLINE __forceinline +#elif defined(__GNUC__) +#define MZ_FORCEINLINE inline __attribute__((__always_inline__)) +#else +#define MZ_FORCEINLINE inline +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +// ------------------- zlib-style API's + +mz_ulong mz_adler32(mz_ulong adler, const unsigned char *ptr, size_t buf_len) { + mz_uint32 i, s1 = (mz_uint32)(adler & 0xffff), s2 = (mz_uint32)(adler >> 16); + size_t block_len = buf_len % 5552; + if (!ptr) return MZ_ADLER32_INIT; + while (buf_len) { + for (i = 0; i + 7 < block_len; i += 8, ptr += 8) { + s1 += ptr[0], s2 += s1; + s1 += ptr[1], s2 += s1; + s1 += ptr[2], s2 += s1; + s1 += ptr[3], s2 += s1; + s1 += ptr[4], s2 += s1; + s1 += ptr[5], s2 += s1; + s1 += ptr[6], s2 += s1; + s1 += ptr[7], s2 += s1; + } + for (; i < block_len; ++i) s1 += *ptr++, s2 += s1; + s1 %= 65521U, s2 %= 65521U; + buf_len -= block_len; + block_len = 5552; + } + return (s2 << 16) + s1; +} + +// Karl Malbrain's compact CRC-32. See "A compact CCITT crc16 and crc32 C +// implementation that balances processor cache usage against speed": +// http://www.geocities.com/malbrain/ +mz_ulong mz_crc32(mz_ulong crc, const mz_uint8 *ptr, size_t buf_len) { + static const mz_uint32 s_crc32[16] = { + 0, 0x1db71064, 0x3b6e20c8, 0x26d930ac, 0x76dc4190, 0x6b6b51f4, + 0x4db26158, 0x5005713c, 0xedb88320, 0xf00f9344, 0xd6d6a3e8, 0xcb61b38c, + 0x9b64c2b0, 0x86d3d2d4, 0xa00ae278, 0xbdbdf21c}; + mz_uint32 crcu32 = (mz_uint32)crc; + if (!ptr) return MZ_CRC32_INIT; + crcu32 = ~crcu32; + while (buf_len--) { + mz_uint8 b = *ptr++; + crcu32 = (crcu32 >> 4) ^ s_crc32[(crcu32 & 0xF) ^ (b & 0xF)]; + crcu32 = (crcu32 >> 4) ^ s_crc32[(crcu32 & 0xF) ^ (b >> 4)]; + } + return ~crcu32; +} + +void mz_free(void *p) { MZ_FREE(p); } + +#ifndef MINIZ_NO_ZLIB_APIS + +static void *def_alloc_func(void *opaque, size_t items, size_t size) { + (void)opaque, (void)items, (void)size; + return MZ_MALLOC(items * size); +} +static void def_free_func(void *opaque, void *address) { + (void)opaque, (void)address; + MZ_FREE(address); +} +// static void *def_realloc_func(void *opaque, void *address, size_t items, +// size_t size) { +// (void)opaque, (void)address, (void)items, (void)size; +// return MZ_REALLOC(address, items * size); +//} + +const char *mz_version(void) { return MZ_VERSION; } + +int mz_deflateInit(mz_streamp pStream, int level) { + return mz_deflateInit2(pStream, level, MZ_DEFLATED, MZ_DEFAULT_WINDOW_BITS, 9, + MZ_DEFAULT_STRATEGY); +} + +int mz_deflateInit2(mz_streamp pStream, int level, int method, int window_bits, + int mem_level, int strategy) { + tdefl_compressor *pComp; + mz_uint comp_flags = + TDEFL_COMPUTE_ADLER32 | + tdefl_create_comp_flags_from_zip_params(level, window_bits, strategy); + + if (!pStream) return MZ_STREAM_ERROR; + if ((method != MZ_DEFLATED) || ((mem_level < 1) || (mem_level > 9)) || + ((window_bits != MZ_DEFAULT_WINDOW_BITS) && + (-window_bits != MZ_DEFAULT_WINDOW_BITS))) + return MZ_PARAM_ERROR; + + pStream->data_type = 0; + pStream->adler = MZ_ADLER32_INIT; + pStream->msg = NULL; + pStream->reserved = 0; + pStream->total_in = 0; + pStream->total_out = 0; + if (!pStream->zalloc) pStream->zalloc = def_alloc_func; + if (!pStream->zfree) pStream->zfree = def_free_func; + + pComp = (tdefl_compressor *)pStream->zalloc(pStream->opaque, 1, + sizeof(tdefl_compressor)); + if (!pComp) return MZ_MEM_ERROR; + + pStream->state = (struct mz_internal_state *)pComp; + + if (tdefl_init(pComp, NULL, NULL, comp_flags) != TDEFL_STATUS_OKAY) { + mz_deflateEnd(pStream); + return MZ_PARAM_ERROR; + } + + return MZ_OK; +} + +int mz_deflateReset(mz_streamp pStream) { + if ((!pStream) || (!pStream->state) || (!pStream->zalloc) || + (!pStream->zfree)) + return MZ_STREAM_ERROR; + pStream->total_in = pStream->total_out = 0; + tdefl_init((tdefl_compressor *)pStream->state, NULL, NULL, + ((tdefl_compressor *)pStream->state)->m_flags); + return MZ_OK; +} + +int mz_deflate(mz_streamp pStream, int flush) { + size_t in_bytes, out_bytes; + mz_ulong orig_total_in, orig_total_out; + int mz_status = MZ_OK; + + if ((!pStream) || (!pStream->state) || (flush < 0) || (flush > MZ_FINISH) || + (!pStream->next_out)) + return MZ_STREAM_ERROR; + if (!pStream->avail_out) return MZ_BUF_ERROR; + + if (flush == MZ_PARTIAL_FLUSH) flush = MZ_SYNC_FLUSH; + + if (((tdefl_compressor *)pStream->state)->m_prev_return_status == + TDEFL_STATUS_DONE) + return (flush == MZ_FINISH) ? MZ_STREAM_END : MZ_BUF_ERROR; + + orig_total_in = pStream->total_in; + orig_total_out = pStream->total_out; + for (;;) { + tdefl_status defl_status; + in_bytes = pStream->avail_in; + out_bytes = pStream->avail_out; + + defl_status = tdefl_compress((tdefl_compressor *)pStream->state, + pStream->next_in, &in_bytes, pStream->next_out, + &out_bytes, (tdefl_flush)flush); + pStream->next_in += (mz_uint)in_bytes; + pStream->avail_in -= (mz_uint)in_bytes; + pStream->total_in += (mz_uint)in_bytes; + pStream->adler = tdefl_get_adler32((tdefl_compressor *)pStream->state); + + pStream->next_out += (mz_uint)out_bytes; + pStream->avail_out -= (mz_uint)out_bytes; + pStream->total_out += (mz_uint)out_bytes; + + if (defl_status < 0) { + mz_status = MZ_STREAM_ERROR; + break; + } else if (defl_status == TDEFL_STATUS_DONE) { + mz_status = MZ_STREAM_END; + break; + } else if (!pStream->avail_out) + break; + else if ((!pStream->avail_in) && (flush != MZ_FINISH)) { + if ((flush) || (pStream->total_in != orig_total_in) || + (pStream->total_out != orig_total_out)) + break; + return MZ_BUF_ERROR; // Can't make forward progress without some input. + } + } + return mz_status; +} + +int mz_deflateEnd(mz_streamp pStream) { + if (!pStream) return MZ_STREAM_ERROR; + if (pStream->state) { + pStream->zfree(pStream->opaque, pStream->state); + pStream->state = NULL; + } + return MZ_OK; +} + +mz_ulong mz_deflateBound(mz_streamp pStream, mz_ulong source_len) { + (void)pStream; + // This is really over conservative. (And lame, but it's actually pretty + // tricky to compute a true upper bound given the way tdefl's blocking works.) + return MZ_MAX(128 + (source_len * 110) / 100, + 128 + source_len + ((source_len / (31 * 1024)) + 1) * 5); +} + +int mz_compress2(unsigned char *pDest, mz_ulong *pDest_len, + const unsigned char *pSource, mz_ulong source_len, int level) { + int status; + mz_stream stream; + memset(&stream, 0, sizeof(stream)); + + // In case mz_ulong is 64-bits (argh I hate longs). + if ((source_len | *pDest_len) > 0xFFFFFFFFU) return MZ_PARAM_ERROR; + + stream.next_in = pSource; + stream.avail_in = (mz_uint32)source_len; + stream.next_out = pDest; + stream.avail_out = (mz_uint32)*pDest_len; + + status = mz_deflateInit(&stream, level); + if (status != MZ_OK) return status; + + status = mz_deflate(&stream, MZ_FINISH); + if (status != MZ_STREAM_END) { + mz_deflateEnd(&stream); + return (status == MZ_OK) ? MZ_BUF_ERROR : status; + } + + *pDest_len = stream.total_out; + return mz_deflateEnd(&stream); +} + +int mz_compress(unsigned char *pDest, mz_ulong *pDest_len, + const unsigned char *pSource, mz_ulong source_len) { + return mz_compress2(pDest, pDest_len, pSource, source_len, + MZ_DEFAULT_COMPRESSION); +} + +mz_ulong mz_compressBound(mz_ulong source_len) { + return mz_deflateBound(NULL, source_len); +} + +typedef struct { + tinfl_decompressor m_decomp; + mz_uint m_dict_ofs, m_dict_avail, m_first_call, m_has_flushed; + int m_window_bits; + mz_uint8 m_dict[TINFL_LZ_DICT_SIZE]; + tinfl_status m_last_status; +} inflate_state; + +int mz_inflateInit2(mz_streamp pStream, int window_bits) { + inflate_state *pDecomp; + if (!pStream) return MZ_STREAM_ERROR; + if ((window_bits != MZ_DEFAULT_WINDOW_BITS) && + (-window_bits != MZ_DEFAULT_WINDOW_BITS)) + return MZ_PARAM_ERROR; + + pStream->data_type = 0; + pStream->adler = 0; + pStream->msg = NULL; + pStream->total_in = 0; + pStream->total_out = 0; + pStream->reserved = 0; + if (!pStream->zalloc) pStream->zalloc = def_alloc_func; + if (!pStream->zfree) pStream->zfree = def_free_func; + + pDecomp = (inflate_state *)pStream->zalloc(pStream->opaque, 1, + sizeof(inflate_state)); + if (!pDecomp) return MZ_MEM_ERROR; + + pStream->state = (struct mz_internal_state *)pDecomp; + + tinfl_init(&pDecomp->m_decomp); + pDecomp->m_dict_ofs = 0; + pDecomp->m_dict_avail = 0; + pDecomp->m_last_status = TINFL_STATUS_NEEDS_MORE_INPUT; + pDecomp->m_first_call = 1; + pDecomp->m_has_flushed = 0; + pDecomp->m_window_bits = window_bits; + + return MZ_OK; +} + +int mz_inflateInit(mz_streamp pStream) { + return mz_inflateInit2(pStream, MZ_DEFAULT_WINDOW_BITS); +} + +int mz_inflate(mz_streamp pStream, int flush) { + inflate_state *pState; + mz_uint n, first_call, decomp_flags = TINFL_FLAG_COMPUTE_ADLER32; + size_t in_bytes, out_bytes, orig_avail_in; + tinfl_status status; + + if ((!pStream) || (!pStream->state)) return MZ_STREAM_ERROR; + if (flush == MZ_PARTIAL_FLUSH) flush = MZ_SYNC_FLUSH; + if ((flush) && (flush != MZ_SYNC_FLUSH) && (flush != MZ_FINISH)) + return MZ_STREAM_ERROR; + + pState = (inflate_state *)pStream->state; + if (pState->m_window_bits > 0) decomp_flags |= TINFL_FLAG_PARSE_ZLIB_HEADER; + orig_avail_in = pStream->avail_in; + + first_call = pState->m_first_call; + pState->m_first_call = 0; + if (pState->m_last_status < 0) return MZ_DATA_ERROR; + + if (pState->m_has_flushed && (flush != MZ_FINISH)) return MZ_STREAM_ERROR; + pState->m_has_flushed |= (flush == MZ_FINISH); + + if ((flush == MZ_FINISH) && (first_call)) { + // MZ_FINISH on the first call implies that the input and output buffers are + // large enough to hold the entire compressed/decompressed file. + decomp_flags |= TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF; + in_bytes = pStream->avail_in; + out_bytes = pStream->avail_out; + status = tinfl_decompress(&pState->m_decomp, pStream->next_in, &in_bytes, + pStream->next_out, pStream->next_out, &out_bytes, + decomp_flags); + pState->m_last_status = status; + pStream->next_in += (mz_uint)in_bytes; + pStream->avail_in -= (mz_uint)in_bytes; + pStream->total_in += (mz_uint)in_bytes; + pStream->adler = tinfl_get_adler32(&pState->m_decomp); + pStream->next_out += (mz_uint)out_bytes; + pStream->avail_out -= (mz_uint)out_bytes; + pStream->total_out += (mz_uint)out_bytes; + + if (status < 0) + return MZ_DATA_ERROR; + else if (status != TINFL_STATUS_DONE) { + pState->m_last_status = TINFL_STATUS_FAILED; + return MZ_BUF_ERROR; + } + return MZ_STREAM_END; + } + // flush != MZ_FINISH then we must assume there's more input. + if (flush != MZ_FINISH) decomp_flags |= TINFL_FLAG_HAS_MORE_INPUT; + + if (pState->m_dict_avail) { + n = MZ_MIN(pState->m_dict_avail, pStream->avail_out); + memcpy(pStream->next_out, pState->m_dict + pState->m_dict_ofs, n); + pStream->next_out += n; + pStream->avail_out -= n; + pStream->total_out += n; + pState->m_dict_avail -= n; + pState->m_dict_ofs = (pState->m_dict_ofs + n) & (TINFL_LZ_DICT_SIZE - 1); + return ((pState->m_last_status == TINFL_STATUS_DONE) && + (!pState->m_dict_avail)) + ? MZ_STREAM_END + : MZ_OK; + } + + for (;;) { + in_bytes = pStream->avail_in; + out_bytes = TINFL_LZ_DICT_SIZE - pState->m_dict_ofs; + + status = tinfl_decompress( + &pState->m_decomp, pStream->next_in, &in_bytes, pState->m_dict, + pState->m_dict + pState->m_dict_ofs, &out_bytes, decomp_flags); + pState->m_last_status = status; + + pStream->next_in += (mz_uint)in_bytes; + pStream->avail_in -= (mz_uint)in_bytes; + pStream->total_in += (mz_uint)in_bytes; + pStream->adler = tinfl_get_adler32(&pState->m_decomp); + + pState->m_dict_avail = (mz_uint)out_bytes; + + n = MZ_MIN(pState->m_dict_avail, pStream->avail_out); + memcpy(pStream->next_out, pState->m_dict + pState->m_dict_ofs, n); + pStream->next_out += n; + pStream->avail_out -= n; + pStream->total_out += n; + pState->m_dict_avail -= n; + pState->m_dict_ofs = (pState->m_dict_ofs + n) & (TINFL_LZ_DICT_SIZE - 1); + + if (status < 0) + return MZ_DATA_ERROR; // Stream is corrupted (there could be some + // uncompressed data left in the output dictionary - + // oh well). + else if ((status == TINFL_STATUS_NEEDS_MORE_INPUT) && (!orig_avail_in)) + return MZ_BUF_ERROR; // Signal caller that we can't make forward progress + // without supplying more input or by setting flush + // to MZ_FINISH. + else if (flush == MZ_FINISH) { + // The output buffer MUST be large to hold the remaining uncompressed data + // when flush==MZ_FINISH. + if (status == TINFL_STATUS_DONE) + return pState->m_dict_avail ? MZ_BUF_ERROR : MZ_STREAM_END; + // status here must be TINFL_STATUS_HAS_MORE_OUTPUT, which means there's + // at least 1 more byte on the way. If there's no more room left in the + // output buffer then something is wrong. + else if (!pStream->avail_out) + return MZ_BUF_ERROR; + } else if ((status == TINFL_STATUS_DONE) || (!pStream->avail_in) || + (!pStream->avail_out) || (pState->m_dict_avail)) + break; + } + + return ((status == TINFL_STATUS_DONE) && (!pState->m_dict_avail)) + ? MZ_STREAM_END + : MZ_OK; +} + +int mz_inflateEnd(mz_streamp pStream) { + if (!pStream) return MZ_STREAM_ERROR; + if (pStream->state) { + pStream->zfree(pStream->opaque, pStream->state); + pStream->state = NULL; + } + return MZ_OK; +} + +int mz_uncompress(unsigned char *pDest, mz_ulong *pDest_len, + const unsigned char *pSource, mz_ulong source_len) { + mz_stream stream; + int status; + memset(&stream, 0, sizeof(stream)); + + // In case mz_ulong is 64-bits (argh I hate longs). + if ((source_len | *pDest_len) > 0xFFFFFFFFU) return MZ_PARAM_ERROR; + + stream.next_in = pSource; + stream.avail_in = (mz_uint32)source_len; + stream.next_out = pDest; + stream.avail_out = (mz_uint32)*pDest_len; + + status = mz_inflateInit(&stream); + if (status != MZ_OK) return status; + + status = mz_inflate(&stream, MZ_FINISH); + if (status != MZ_STREAM_END) { + mz_inflateEnd(&stream); + return ((status == MZ_BUF_ERROR) && (!stream.avail_in)) ? MZ_DATA_ERROR + : status; + } + *pDest_len = stream.total_out; + + return mz_inflateEnd(&stream); +} + +const char *mz_error(int err) { + static struct { + int m_err; + const char *m_pDesc; + } s_error_descs[] = {{MZ_OK, ""}, + {MZ_STREAM_END, "stream end"}, + {MZ_NEED_DICT, "need dictionary"}, + {MZ_ERRNO, "file error"}, + {MZ_STREAM_ERROR, "stream error"}, + {MZ_DATA_ERROR, "data error"}, + {MZ_MEM_ERROR, "out of memory"}, + {MZ_BUF_ERROR, "buf error"}, + {MZ_VERSION_ERROR, "version error"}, + {MZ_PARAM_ERROR, "parameter error"}}; + mz_uint i; + for (i = 0; i < sizeof(s_error_descs) / sizeof(s_error_descs[0]); ++i) + if (s_error_descs[i].m_err == err) return s_error_descs[i].m_pDesc; + return NULL; +} + +#endif // MINIZ_NO_ZLIB_APIS + +// ------------------- Low-level Decompression (completely independent from all +// compression API's) + +#define TINFL_MEMCPY(d, s, l) memcpy(d, s, l) +#define TINFL_MEMSET(p, c, l) memset(p, c, l) + +#define TINFL_CR_BEGIN \ + switch (r->m_state) { \ + case 0: +#define TINFL_CR_RETURN(state_index, result) \ + do { \ + status = result; \ + r->m_state = state_index; \ + goto common_exit; \ + case state_index:; \ + } \ + MZ_MACRO_END +#define TINFL_CR_RETURN_FOREVER(state_index, result) \ + do { \ + for (;;) { \ + TINFL_CR_RETURN(state_index, result); \ + } \ + } \ + MZ_MACRO_END +#define TINFL_CR_FINISH } + +// TODO: If the caller has indicated that there's no more input, and we attempt +// to read beyond the input buf, then something is wrong with the input because +// the inflator never +// reads ahead more than it needs to. Currently TINFL_GET_BYTE() pads the end of +// the stream with 0's in this scenario. +#define TINFL_GET_BYTE(state_index, c) \ + do { \ + if (pIn_buf_cur >= pIn_buf_end) { \ + for (;;) { \ + if (decomp_flags & TINFL_FLAG_HAS_MORE_INPUT) { \ + TINFL_CR_RETURN(state_index, TINFL_STATUS_NEEDS_MORE_INPUT); \ + if (pIn_buf_cur < pIn_buf_end) { \ + c = *pIn_buf_cur++; \ + break; \ + } \ + } else { \ + c = 0; \ + break; \ + } \ + } \ + } else \ + c = *pIn_buf_cur++; \ + } \ + MZ_MACRO_END + +#define TINFL_NEED_BITS(state_index, n) \ + do { \ + mz_uint c; \ + TINFL_GET_BYTE(state_index, c); \ + bit_buf |= (((tinfl_bit_buf_t)c) << num_bits); \ + num_bits += 8; \ + } while (num_bits < (mz_uint)(n)) +#define TINFL_SKIP_BITS(state_index, n) \ + do { \ + if (num_bits < (mz_uint)(n)) { \ + TINFL_NEED_BITS(state_index, n); \ + } \ + bit_buf >>= (n); \ + num_bits -= (n); \ + } \ + MZ_MACRO_END +#define TINFL_GET_BITS(state_index, b, n) \ + do { \ + if (num_bits < (mz_uint)(n)) { \ + TINFL_NEED_BITS(state_index, n); \ + } \ + b = bit_buf & ((1 << (n)) - 1); \ + bit_buf >>= (n); \ + num_bits -= (n); \ + } \ + MZ_MACRO_END + +// TINFL_HUFF_BITBUF_FILL() is only used rarely, when the number of bytes +// remaining in the input buffer falls below 2. +// It reads just enough bytes from the input stream that are needed to decode +// the next Huffman code (and absolutely no more). It works by trying to fully +// decode a +// Huffman code by using whatever bits are currently present in the bit buffer. +// If this fails, it reads another byte, and tries again until it succeeds or +// until the +// bit buffer contains >=15 bits (deflate's max. Huffman code size). +#define TINFL_HUFF_BITBUF_FILL(state_index, pHuff) \ + do { \ + temp = (pHuff)->m_look_up[bit_buf & (TINFL_FAST_LOOKUP_SIZE - 1)]; \ + if (temp >= 0) { \ + code_len = temp >> 9; \ + if ((code_len) && (num_bits >= code_len)) break; \ + } else if (num_bits > TINFL_FAST_LOOKUP_BITS) { \ + code_len = TINFL_FAST_LOOKUP_BITS; \ + do { \ + temp = (pHuff)->m_tree[~temp + ((bit_buf >> code_len++) & 1)]; \ + } while ((temp < 0) && (num_bits >= (code_len + 1))); \ + if (temp >= 0) break; \ + } \ + TINFL_GET_BYTE(state_index, c); \ + bit_buf |= (((tinfl_bit_buf_t)c) << num_bits); \ + num_bits += 8; \ + } while (num_bits < 15); + +// TINFL_HUFF_DECODE() decodes the next Huffman coded symbol. It's more complex +// than you would initially expect because the zlib API expects the decompressor +// to never read +// beyond the final byte of the deflate stream. (In other words, when this macro +// wants to read another byte from the input, it REALLY needs another byte in +// order to fully +// decode the next Huffman code.) Handling this properly is particularly +// important on raw deflate (non-zlib) streams, which aren't followed by a byte +// aligned adler-32. +// The slow path is only executed at the very end of the input buffer. +#define TINFL_HUFF_DECODE(state_index, sym, pHuff) \ + do { \ + int temp; \ + mz_uint code_len, c; \ + if (num_bits < 15) { \ + if ((pIn_buf_end - pIn_buf_cur) < 2) { \ + TINFL_HUFF_BITBUF_FILL(state_index, pHuff); \ + } else { \ + bit_buf |= (((tinfl_bit_buf_t)pIn_buf_cur[0]) << num_bits) | \ + (((tinfl_bit_buf_t)pIn_buf_cur[1]) << (num_bits + 8)); \ + pIn_buf_cur += 2; \ + num_bits += 16; \ + } \ + } \ + if ((temp = (pHuff)->m_look_up[bit_buf & (TINFL_FAST_LOOKUP_SIZE - 1)]) >= \ + 0) \ + code_len = temp >> 9, temp &= 511; \ + else { \ + code_len = TINFL_FAST_LOOKUP_BITS; \ + do { \ + temp = (pHuff)->m_tree[~temp + ((bit_buf >> code_len++) & 1)]; \ + } while (temp < 0); \ + } \ + sym = temp; \ + bit_buf >>= code_len; \ + num_bits -= code_len; \ + } \ + MZ_MACRO_END + +tinfl_status tinfl_decompress(tinfl_decompressor *r, + const mz_uint8 *pIn_buf_next, + size_t *pIn_buf_size, mz_uint8 *pOut_buf_start, + mz_uint8 *pOut_buf_next, size_t *pOut_buf_size, + const mz_uint32 decomp_flags) { + static const int s_length_base[31] = { + 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31, + 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0}; + static const int s_length_extra[31] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, + 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, + 4, 4, 5, 5, 5, 5, 0, 0, 0}; + static const int s_dist_base[32] = { + 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, + 49, 65, 97, 129, 193, 257, 385, 513, 769, 1025, 1537, + 2049, 3073, 4097, 6145, 8193, 12289, 16385, 24577, 0, 0}; + static const int s_dist_extra[32] = {0, 0, 0, 0, 1, 1, 2, 2, 3, 3, + 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, + 9, 9, 10, 10, 11, 11, 12, 12, 13, 13}; + static const mz_uint8 s_length_dezigzag[19] = { + 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15}; + static const int s_min_table_sizes[3] = {257, 1, 4}; + + tinfl_status status = TINFL_STATUS_FAILED; + mz_uint32 num_bits, dist, counter, num_extra; + tinfl_bit_buf_t bit_buf; + const mz_uint8 *pIn_buf_cur = pIn_buf_next, *const pIn_buf_end = + pIn_buf_next + *pIn_buf_size; + mz_uint8 *pOut_buf_cur = pOut_buf_next, *const pOut_buf_end = + pOut_buf_next + *pOut_buf_size; + size_t out_buf_size_mask = + (decomp_flags & TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF) + ? (size_t)-1 + : ((pOut_buf_next - pOut_buf_start) + *pOut_buf_size) - 1, + dist_from_out_buf_start; + + // Ensure the output buffer's size is a power of 2, unless the output buffer + // is large enough to hold the entire output file (in which case it doesn't + // matter). + if (((out_buf_size_mask + 1) & out_buf_size_mask) || + (pOut_buf_next < pOut_buf_start)) { + *pIn_buf_size = *pOut_buf_size = 0; + return TINFL_STATUS_BAD_PARAM; + } + + num_bits = r->m_num_bits; + bit_buf = r->m_bit_buf; + dist = r->m_dist; + counter = r->m_counter; + num_extra = r->m_num_extra; + dist_from_out_buf_start = r->m_dist_from_out_buf_start; + TINFL_CR_BEGIN + + bit_buf = num_bits = dist = counter = num_extra = r->m_zhdr0 = r->m_zhdr1 = 0; + r->m_z_adler32 = r->m_check_adler32 = 1; + if (decomp_flags & TINFL_FLAG_PARSE_ZLIB_HEADER) { + TINFL_GET_BYTE(1, r->m_zhdr0); + TINFL_GET_BYTE(2, r->m_zhdr1); + counter = (((r->m_zhdr0 * 256 + r->m_zhdr1) % 31 != 0) || + (r->m_zhdr1 & 32) || ((r->m_zhdr0 & 15) != 8)); + if (!(decomp_flags & TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF)) + counter |= (((1U << (8U + (r->m_zhdr0 >> 4))) > 32768U) || + ((out_buf_size_mask + 1) < + (size_t)(1ULL << (8U + (r->m_zhdr0 >> 4))))); + if (counter) { + TINFL_CR_RETURN_FOREVER(36, TINFL_STATUS_FAILED); + } + } + + do { + TINFL_GET_BITS(3, r->m_final, 3); + r->m_type = r->m_final >> 1; + if (r->m_type == 0) { + TINFL_SKIP_BITS(5, num_bits & 7); + for (counter = 0; counter < 4; ++counter) { + if (num_bits) + TINFL_GET_BITS(6, r->m_raw_header[counter], 8); + else + TINFL_GET_BYTE(7, r->m_raw_header[counter]); + } + if ((counter = (r->m_raw_header[0] | (r->m_raw_header[1] << 8))) != + (mz_uint)(0xFFFF ^ + (r->m_raw_header[2] | (r->m_raw_header[3] << 8)))) { + TINFL_CR_RETURN_FOREVER(39, TINFL_STATUS_FAILED); + } + while ((counter) && (num_bits)) { + TINFL_GET_BITS(51, dist, 8); + while (pOut_buf_cur >= pOut_buf_end) { + TINFL_CR_RETURN(52, TINFL_STATUS_HAS_MORE_OUTPUT); + } + *pOut_buf_cur++ = (mz_uint8)dist; + counter--; + } + while (counter) { + size_t n; + while (pOut_buf_cur >= pOut_buf_end) { + TINFL_CR_RETURN(9, TINFL_STATUS_HAS_MORE_OUTPUT); + } + while (pIn_buf_cur >= pIn_buf_end) { + if (decomp_flags & TINFL_FLAG_HAS_MORE_INPUT) { + TINFL_CR_RETURN(38, TINFL_STATUS_NEEDS_MORE_INPUT); + } else { + TINFL_CR_RETURN_FOREVER(40, TINFL_STATUS_FAILED); + } + } + n = MZ_MIN(MZ_MIN((size_t)(pOut_buf_end - pOut_buf_cur), + (size_t)(pIn_buf_end - pIn_buf_cur)), + counter); + TINFL_MEMCPY(pOut_buf_cur, pIn_buf_cur, n); + pIn_buf_cur += n; + pOut_buf_cur += n; + counter -= (mz_uint)n; + } + } else if (r->m_type == 3) { + TINFL_CR_RETURN_FOREVER(10, TINFL_STATUS_FAILED); + } else { + if (r->m_type == 1) { + mz_uint8 *p = r->m_tables[0].m_code_size; + mz_uint i; + r->m_table_sizes[0] = 288; + r->m_table_sizes[1] = 32; + TINFL_MEMSET(r->m_tables[1].m_code_size, 5, 32); + for (i = 0; i <= 143; ++i) *p++ = 8; + for (; i <= 255; ++i) *p++ = 9; + for (; i <= 279; ++i) *p++ = 7; + for (; i <= 287; ++i) *p++ = 8; + } else { + for (counter = 0; counter < 3; counter++) { + TINFL_GET_BITS(11, r->m_table_sizes[counter], "\05\05\04"[counter]); + r->m_table_sizes[counter] += s_min_table_sizes[counter]; + } + MZ_CLEAR_OBJ(r->m_tables[2].m_code_size); + for (counter = 0; counter < r->m_table_sizes[2]; counter++) { + mz_uint s; + TINFL_GET_BITS(14, s, 3); + r->m_tables[2].m_code_size[s_length_dezigzag[counter]] = (mz_uint8)s; + } + r->m_table_sizes[2] = 19; + } + for (; (int)r->m_type >= 0; r->m_type--) { + int tree_next, tree_cur; + tinfl_huff_table *pTable; + mz_uint i, j, used_syms, total, sym_index, next_code[17], + total_syms[16]; + pTable = &r->m_tables[r->m_type]; + MZ_CLEAR_OBJ(total_syms); + MZ_CLEAR_OBJ(pTable->m_look_up); + MZ_CLEAR_OBJ(pTable->m_tree); + for (i = 0; i < r->m_table_sizes[r->m_type]; ++i) + total_syms[pTable->m_code_size[i]]++; + used_syms = 0, total = 0; + next_code[0] = next_code[1] = 0; + for (i = 1; i <= 15; ++i) { + used_syms += total_syms[i]; + next_code[i + 1] = (total = ((total + total_syms[i]) << 1)); + } + if ((65536 != total) && (used_syms > 1)) { + TINFL_CR_RETURN_FOREVER(35, TINFL_STATUS_FAILED); + } + for (tree_next = -1, sym_index = 0; + sym_index < r->m_table_sizes[r->m_type]; ++sym_index) { + mz_uint rev_code = 0, l, cur_code, + code_size = pTable->m_code_size[sym_index]; + if (!code_size) continue; + cur_code = next_code[code_size]++; + for (l = code_size; l > 0; l--, cur_code >>= 1) + rev_code = (rev_code << 1) | (cur_code & 1); + if (code_size <= TINFL_FAST_LOOKUP_BITS) { + mz_int16 k = (mz_int16)((code_size << 9) | sym_index); + while (rev_code < TINFL_FAST_LOOKUP_SIZE) { + pTable->m_look_up[rev_code] = k; + rev_code += (1 << code_size); + } + continue; + } + if (0 == + (tree_cur = pTable->m_look_up[rev_code & + (TINFL_FAST_LOOKUP_SIZE - 1)])) { + pTable->m_look_up[rev_code & (TINFL_FAST_LOOKUP_SIZE - 1)] = + (mz_int16)tree_next; + tree_cur = tree_next; + tree_next -= 2; + } + rev_code >>= (TINFL_FAST_LOOKUP_BITS - 1); + for (j = code_size; j > (TINFL_FAST_LOOKUP_BITS + 1); j--) { + tree_cur -= ((rev_code >>= 1) & 1); + if (!pTable->m_tree[-tree_cur - 1]) { + pTable->m_tree[-tree_cur - 1] = (mz_int16)tree_next; + tree_cur = tree_next; + tree_next -= 2; + } else + tree_cur = pTable->m_tree[-tree_cur - 1]; + } + tree_cur -= ((rev_code >>= 1) & 1); + pTable->m_tree[-tree_cur - 1] = (mz_int16)sym_index; + } + if (r->m_type == 2) { + for (counter = 0; + counter < (r->m_table_sizes[0] + r->m_table_sizes[1]);) { + mz_uint s; + TINFL_HUFF_DECODE(16, dist, &r->m_tables[2]); + if (dist < 16) { + r->m_len_codes[counter++] = (mz_uint8)dist; + continue; + } + if ((dist == 16) && (!counter)) { + TINFL_CR_RETURN_FOREVER(17, TINFL_STATUS_FAILED); + } + num_extra = "\02\03\07"[dist - 16]; + TINFL_GET_BITS(18, s, num_extra); + s += "\03\03\013"[dist - 16]; + TINFL_MEMSET(r->m_len_codes + counter, + (dist == 16) ? r->m_len_codes[counter - 1] : 0, s); + counter += s; + } + if ((r->m_table_sizes[0] + r->m_table_sizes[1]) != counter) { + TINFL_CR_RETURN_FOREVER(21, TINFL_STATUS_FAILED); + } + TINFL_MEMCPY(r->m_tables[0].m_code_size, r->m_len_codes, + r->m_table_sizes[0]); + TINFL_MEMCPY(r->m_tables[1].m_code_size, + r->m_len_codes + r->m_table_sizes[0], + r->m_table_sizes[1]); + } + } + for (;;) { + mz_uint8 *pSrc; + for (;;) { + if (((pIn_buf_end - pIn_buf_cur) < 4) || + ((pOut_buf_end - pOut_buf_cur) < 2)) { + TINFL_HUFF_DECODE(23, counter, &r->m_tables[0]); + if (counter >= 256) break; + while (pOut_buf_cur >= pOut_buf_end) { + TINFL_CR_RETURN(24, TINFL_STATUS_HAS_MORE_OUTPUT); + } + *pOut_buf_cur++ = (mz_uint8)counter; + } else { + int sym2; + mz_uint code_len; +#if TINFL_USE_64BIT_BITBUF + if (num_bits < 30) { + bit_buf |= + (((tinfl_bit_buf_t)MZ_READ_LE32(pIn_buf_cur)) << num_bits); + pIn_buf_cur += 4; + num_bits += 32; + } +#else + if (num_bits < 15) { + bit_buf |= + (((tinfl_bit_buf_t)MZ_READ_LE16(pIn_buf_cur)) << num_bits); + pIn_buf_cur += 2; + num_bits += 16; + } +#endif + if ((sym2 = + r->m_tables[0] + .m_look_up[bit_buf & (TINFL_FAST_LOOKUP_SIZE - 1)]) >= + 0) + code_len = sym2 >> 9; + else { + code_len = TINFL_FAST_LOOKUP_BITS; + do { + sym2 = r->m_tables[0] + .m_tree[~sym2 + ((bit_buf >> code_len++) & 1)]; + } while (sym2 < 0); + } + counter = sym2; + bit_buf >>= code_len; + num_bits -= code_len; + if (counter & 256) break; + +#if !TINFL_USE_64BIT_BITBUF + if (num_bits < 15) { + bit_buf |= + (((tinfl_bit_buf_t)MZ_READ_LE16(pIn_buf_cur)) << num_bits); + pIn_buf_cur += 2; + num_bits += 16; + } +#endif + if ((sym2 = + r->m_tables[0] + .m_look_up[bit_buf & (TINFL_FAST_LOOKUP_SIZE - 1)]) >= + 0) + code_len = sym2 >> 9; + else { + code_len = TINFL_FAST_LOOKUP_BITS; + do { + sym2 = r->m_tables[0] + .m_tree[~sym2 + ((bit_buf >> code_len++) & 1)]; + } while (sym2 < 0); + } + bit_buf >>= code_len; + num_bits -= code_len; + + pOut_buf_cur[0] = (mz_uint8)counter; + if (sym2 & 256) { + pOut_buf_cur++; + counter = sym2; + break; + } + pOut_buf_cur[1] = (mz_uint8)sym2; + pOut_buf_cur += 2; + } + } + if ((counter &= 511) == 256) break; + + num_extra = s_length_extra[counter - 257]; + counter = s_length_base[counter - 257]; + if (num_extra) { + mz_uint extra_bits; + TINFL_GET_BITS(25, extra_bits, num_extra); + counter += extra_bits; + } + + TINFL_HUFF_DECODE(26, dist, &r->m_tables[1]); + num_extra = s_dist_extra[dist]; + dist = s_dist_base[dist]; + if (num_extra) { + mz_uint extra_bits; + TINFL_GET_BITS(27, extra_bits, num_extra); + dist += extra_bits; + } + + dist_from_out_buf_start = pOut_buf_cur - pOut_buf_start; + if ((dist > dist_from_out_buf_start) && + (decomp_flags & TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF)) { + TINFL_CR_RETURN_FOREVER(37, TINFL_STATUS_FAILED); + } + + pSrc = pOut_buf_start + + ((dist_from_out_buf_start - dist) & out_buf_size_mask); + + if ((MZ_MAX(pOut_buf_cur, pSrc) + counter) > pOut_buf_end) { + while (counter--) { + while (pOut_buf_cur >= pOut_buf_end) { + TINFL_CR_RETURN(53, TINFL_STATUS_HAS_MORE_OUTPUT); + } + *pOut_buf_cur++ = + pOut_buf_start[(dist_from_out_buf_start++ - dist) & + out_buf_size_mask]; + } + continue; + } +#if MINIZ_USE_UNALIGNED_LOADS_AND_STORES + else if ((counter >= 9) && (counter <= dist)) { + const mz_uint8 *pSrc_end = pSrc + (counter & ~7); + do { + ((mz_uint32 *)pOut_buf_cur)[0] = ((const mz_uint32 *)pSrc)[0]; + ((mz_uint32 *)pOut_buf_cur)[1] = ((const mz_uint32 *)pSrc)[1]; + pOut_buf_cur += 8; + } while ((pSrc += 8) < pSrc_end); + if ((counter &= 7) < 3) { + if (counter) { + pOut_buf_cur[0] = pSrc[0]; + if (counter > 1) pOut_buf_cur[1] = pSrc[1]; + pOut_buf_cur += counter; + } + continue; + } + } +#endif + do { + pOut_buf_cur[0] = pSrc[0]; + pOut_buf_cur[1] = pSrc[1]; + pOut_buf_cur[2] = pSrc[2]; + pOut_buf_cur += 3; + pSrc += 3; + } while ((int)(counter -= 3) > 2); + if ((int)counter > 0) { + pOut_buf_cur[0] = pSrc[0]; + if ((int)counter > 1) pOut_buf_cur[1] = pSrc[1]; + pOut_buf_cur += counter; + } + } + } + } while (!(r->m_final & 1)); + if (decomp_flags & TINFL_FLAG_PARSE_ZLIB_HEADER) { + TINFL_SKIP_BITS(32, num_bits & 7); + for (counter = 0; counter < 4; ++counter) { + mz_uint s; + if (num_bits) + TINFL_GET_BITS(41, s, 8); + else + TINFL_GET_BYTE(42, s); + r->m_z_adler32 = (r->m_z_adler32 << 8) | s; + } + } + TINFL_CR_RETURN_FOREVER(34, TINFL_STATUS_DONE); + TINFL_CR_FINISH + +common_exit: + r->m_num_bits = num_bits; + r->m_bit_buf = bit_buf; + r->m_dist = dist; + r->m_counter = counter; + r->m_num_extra = num_extra; + r->m_dist_from_out_buf_start = dist_from_out_buf_start; + *pIn_buf_size = pIn_buf_cur - pIn_buf_next; + *pOut_buf_size = pOut_buf_cur - pOut_buf_next; + if ((decomp_flags & + (TINFL_FLAG_PARSE_ZLIB_HEADER | TINFL_FLAG_COMPUTE_ADLER32)) && + (status >= 0)) { + const mz_uint8 *ptr = pOut_buf_next; + size_t buf_len = *pOut_buf_size; + mz_uint32 i, s1 = r->m_check_adler32 & 0xffff, + s2 = r->m_check_adler32 >> 16; + size_t block_len = buf_len % 5552; + while (buf_len) { + for (i = 0; i + 7 < block_len; i += 8, ptr += 8) { + s1 += ptr[0], s2 += s1; + s1 += ptr[1], s2 += s1; + s1 += ptr[2], s2 += s1; + s1 += ptr[3], s2 += s1; + s1 += ptr[4], s2 += s1; + s1 += ptr[5], s2 += s1; + s1 += ptr[6], s2 += s1; + s1 += ptr[7], s2 += s1; + } + for (; i < block_len; ++i) s1 += *ptr++, s2 += s1; + s1 %= 65521U, s2 %= 65521U; + buf_len -= block_len; + block_len = 5552; + } + r->m_check_adler32 = (s2 << 16) + s1; + if ((status == TINFL_STATUS_DONE) && + (decomp_flags & TINFL_FLAG_PARSE_ZLIB_HEADER) && + (r->m_check_adler32 != r->m_z_adler32)) + status = TINFL_STATUS_ADLER32_MISMATCH; + } + return status; +} + +// Higher level helper functions. +void *tinfl_decompress_mem_to_heap(const void *pSrc_buf, size_t src_buf_len, + size_t *pOut_len, int flags) { + tinfl_decompressor decomp; + void *pBuf = NULL, *pNew_buf; + size_t src_buf_ofs = 0, out_buf_capacity = 0; + *pOut_len = 0; + tinfl_init(&decomp); + for (;;) { + size_t src_buf_size = src_buf_len - src_buf_ofs, + dst_buf_size = out_buf_capacity - *pOut_len, new_out_buf_capacity; + tinfl_status status = tinfl_decompress( + &decomp, (const mz_uint8 *)pSrc_buf + src_buf_ofs, &src_buf_size, + (mz_uint8 *)pBuf, pBuf ? (mz_uint8 *)pBuf + *pOut_len : NULL, + &dst_buf_size, + (flags & ~TINFL_FLAG_HAS_MORE_INPUT) | + TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF); + if ((status < 0) || (status == TINFL_STATUS_NEEDS_MORE_INPUT)) { + MZ_FREE(pBuf); + *pOut_len = 0; + return NULL; + } + src_buf_ofs += src_buf_size; + *pOut_len += dst_buf_size; + if (status == TINFL_STATUS_DONE) break; + new_out_buf_capacity = out_buf_capacity * 2; + if (new_out_buf_capacity < 128) new_out_buf_capacity = 128; + pNew_buf = MZ_REALLOC(pBuf, new_out_buf_capacity); + if (!pNew_buf) { + MZ_FREE(pBuf); + *pOut_len = 0; + return NULL; + } + pBuf = pNew_buf; + out_buf_capacity = new_out_buf_capacity; + } + return pBuf; +} + +size_t tinfl_decompress_mem_to_mem(void *pOut_buf, size_t out_buf_len, + const void *pSrc_buf, size_t src_buf_len, + int flags) { + tinfl_decompressor decomp; + tinfl_status status; + tinfl_init(&decomp); + status = + tinfl_decompress(&decomp, (const mz_uint8 *)pSrc_buf, &src_buf_len, + (mz_uint8 *)pOut_buf, (mz_uint8 *)pOut_buf, &out_buf_len, + (flags & ~TINFL_FLAG_HAS_MORE_INPUT) | + TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF); + return (status != TINFL_STATUS_DONE) ? TINFL_DECOMPRESS_MEM_TO_MEM_FAILED + : out_buf_len; +} + +int tinfl_decompress_mem_to_callback(const void *pIn_buf, size_t *pIn_buf_size, + tinfl_put_buf_func_ptr pPut_buf_func, + void *pPut_buf_user, int flags) { + int result = 0; + tinfl_decompressor decomp; + mz_uint8 *pDict = (mz_uint8 *)MZ_MALLOC(TINFL_LZ_DICT_SIZE); + size_t in_buf_ofs = 0, dict_ofs = 0; + if (!pDict) return TINFL_STATUS_FAILED; + tinfl_init(&decomp); + for (;;) { + size_t in_buf_size = *pIn_buf_size - in_buf_ofs, + dst_buf_size = TINFL_LZ_DICT_SIZE - dict_ofs; + tinfl_status status = + tinfl_decompress(&decomp, (const mz_uint8 *)pIn_buf + in_buf_ofs, + &in_buf_size, pDict, pDict + dict_ofs, &dst_buf_size, + (flags & ~(TINFL_FLAG_HAS_MORE_INPUT | + TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF))); + in_buf_ofs += in_buf_size; + if ((dst_buf_size) && + (!(*pPut_buf_func)(pDict + dict_ofs, (int)dst_buf_size, pPut_buf_user))) + break; + if (status != TINFL_STATUS_HAS_MORE_OUTPUT) { + result = (status == TINFL_STATUS_DONE); + break; + } + dict_ofs = (dict_ofs + dst_buf_size) & (TINFL_LZ_DICT_SIZE - 1); + } + MZ_FREE(pDict); + *pIn_buf_size = in_buf_ofs; + return result; +} + +// ------------------- Low-level Compression (independent from all decompression +// API's) + +// Purposely making these tables static for faster init and thread safety. +static const mz_uint16 s_tdefl_len_sym[256] = { + 257, 258, 259, 260, 261, 262, 263, 264, 265, 265, 266, 266, 267, 267, 268, + 268, 269, 269, 269, 269, 270, 270, 270, 270, 271, 271, 271, 271, 272, 272, + 272, 272, 273, 273, 273, 273, 273, 273, 273, 273, 274, 274, 274, 274, 274, + 274, 274, 274, 275, 275, 275, 275, 275, 275, 275, 275, 276, 276, 276, 276, + 276, 276, 276, 276, 277, 277, 277, 277, 277, 277, 277, 277, 277, 277, 277, + 277, 277, 277, 277, 277, 278, 278, 278, 278, 278, 278, 278, 278, 278, 278, + 278, 278, 278, 278, 278, 278, 279, 279, 279, 279, 279, 279, 279, 279, 279, + 279, 279, 279, 279, 279, 279, 279, 280, 280, 280, 280, 280, 280, 280, 280, + 280, 280, 280, 280, 280, 280, 280, 280, 281, 281, 281, 281, 281, 281, 281, + 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, + 281, 281, 281, 281, 281, 281, 281, 281, 281, 281, 282, 282, 282, 282, 282, + 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, + 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 282, 283, 283, 283, + 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, + 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 284, + 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, + 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, 284, + 285}; + +static const mz_uint8 s_tdefl_len_extra[256] = { + 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, + 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, + 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, + 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 0}; + +static const mz_uint8 s_tdefl_small_dist_sym[512] = { + 0, 1, 2, 3, 4, 4, 5, 5, 6, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, + 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 10, + 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 11, + 11, 11, 11, 11, 11, 11, 11, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, + 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, + 12, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 14, 14, 14, 14, 14, + 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, + 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, + 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, + 14, 14, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, + 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, + 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, + 15, 15, 15, 15, 15, 15, 15, 15, 15, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17}; + +static const mz_uint8 s_tdefl_small_dist_extra[512] = { + 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, + 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7}; + +static const mz_uint8 s_tdefl_large_dist_sym[128] = { + 0, 0, 18, 19, 20, 20, 21, 21, 22, 22, 22, 22, 23, 23, 23, 23, 24, 24, 24, + 24, 24, 24, 24, 24, 25, 25, 25, 25, 25, 25, 25, 25, 26, 26, 26, 26, 26, 26, + 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 27, 27, 27, 27, 27, 27, 27, 27, 27, + 27, 27, 27, 27, 27, 27, 27, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, + 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, + 28, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, + 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29}; + +static const mz_uint8 s_tdefl_large_dist_extra[128] = { + 0, 0, 8, 8, 9, 9, 9, 9, 10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, + 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 12, 12, 12, 12, 12, 12, + 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, + 12, 12, 12, 12, 12, 12, 12, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13}; + +// Radix sorts tdefl_sym_freq[] array by 16-bit key m_key. Returns ptr to sorted +// values. +typedef struct { + mz_uint16 m_key, m_sym_index; +} tdefl_sym_freq; +static tdefl_sym_freq *tdefl_radix_sort_syms(mz_uint num_syms, + tdefl_sym_freq *pSyms0, + tdefl_sym_freq *pSyms1) { + mz_uint32 total_passes = 2, pass_shift, pass, i, hist[256 * 2]; + tdefl_sym_freq *pCur_syms = pSyms0, *pNew_syms = pSyms1; + MZ_CLEAR_OBJ(hist); + for (i = 0; i < num_syms; i++) { + mz_uint freq = pSyms0[i].m_key; + hist[freq & 0xFF]++; + hist[256 + ((freq >> 8) & 0xFF)]++; + } + while ((total_passes > 1) && (num_syms == hist[(total_passes - 1) * 256])) + total_passes--; + for (pass_shift = 0, pass = 0; pass < total_passes; pass++, pass_shift += 8) { + const mz_uint32 *pHist = &hist[pass << 8]; + mz_uint offsets[256], cur_ofs = 0; + for (i = 0; i < 256; i++) { + offsets[i] = cur_ofs; + cur_ofs += pHist[i]; + } + for (i = 0; i < num_syms; i++) + pNew_syms[offsets[(pCur_syms[i].m_key >> pass_shift) & 0xFF]++] = + pCur_syms[i]; + { + tdefl_sym_freq *t = pCur_syms; + pCur_syms = pNew_syms; + pNew_syms = t; + } + } + return pCur_syms; +} + +// tdefl_calculate_minimum_redundancy() originally written by: Alistair Moffat, +// alistair@cs.mu.oz.au, Jyrki Katajainen, jyrki@diku.dk, November 1996. +static void tdefl_calculate_minimum_redundancy(tdefl_sym_freq *A, int n) { + int root, leaf, next, avbl, used, dpth; + if (n == 0) + return; + else if (n == 1) { + A[0].m_key = 1; + return; + } + A[0].m_key += A[1].m_key; + root = 0; + leaf = 2; + for (next = 1; next < n - 1; next++) { + if (leaf >= n || A[root].m_key < A[leaf].m_key) { + A[next].m_key = A[root].m_key; + A[root++].m_key = (mz_uint16)next; + } else + A[next].m_key = A[leaf++].m_key; + if (leaf >= n || (root < next && A[root].m_key < A[leaf].m_key)) { + A[next].m_key = (mz_uint16)(A[next].m_key + A[root].m_key); + A[root++].m_key = (mz_uint16)next; + } else + A[next].m_key = (mz_uint16)(A[next].m_key + A[leaf++].m_key); + } + A[n - 2].m_key = 0; + for (next = n - 3; next >= 0; next--) + A[next].m_key = A[A[next].m_key].m_key + 1; + avbl = 1; + used = dpth = 0; + root = n - 2; + next = n - 1; + while (avbl > 0) { + while (root >= 0 && (int)A[root].m_key == dpth) { + used++; + root--; + } + while (avbl > used) { + A[next--].m_key = (mz_uint16)(dpth); + avbl--; + } + avbl = 2 * used; + dpth++; + used = 0; + } +} + +// Limits canonical Huffman code table's max code size. +enum { TDEFL_MAX_SUPPORTED_HUFF_CODESIZE = 32 }; +static void tdefl_huffman_enforce_max_code_size(int *pNum_codes, + int code_list_len, + int max_code_size) { + int i; + mz_uint32 total = 0; + if (code_list_len <= 1) return; + for (i = max_code_size + 1; i <= TDEFL_MAX_SUPPORTED_HUFF_CODESIZE; i++) + pNum_codes[max_code_size] += pNum_codes[i]; + for (i = max_code_size; i > 0; i--) + total += (((mz_uint32)pNum_codes[i]) << (max_code_size - i)); + while (total != (1UL << max_code_size)) { + pNum_codes[max_code_size]--; + for (i = max_code_size - 1; i > 0; i--) + if (pNum_codes[i]) { + pNum_codes[i]--; + pNum_codes[i + 1] += 2; + break; + } + total--; + } +} + +static void tdefl_optimize_huffman_table(tdefl_compressor *d, int table_num, + int table_len, int code_size_limit, + int static_table) { + int i, j, l, num_codes[1 + TDEFL_MAX_SUPPORTED_HUFF_CODESIZE]; + mz_uint next_code[TDEFL_MAX_SUPPORTED_HUFF_CODESIZE + 1]; + MZ_CLEAR_OBJ(num_codes); + if (static_table) { + for (i = 0; i < table_len; i++) + num_codes[d->m_huff_code_sizes[table_num][i]]++; + } else { + tdefl_sym_freq syms0[TDEFL_MAX_HUFF_SYMBOLS], syms1[TDEFL_MAX_HUFF_SYMBOLS], + *pSyms; + int num_used_syms = 0; + const mz_uint16 *pSym_count = &d->m_huff_count[table_num][0]; + for (i = 0; i < table_len; i++) + if (pSym_count[i]) { + syms0[num_used_syms].m_key = (mz_uint16)pSym_count[i]; + syms0[num_used_syms++].m_sym_index = (mz_uint16)i; + } + + pSyms = tdefl_radix_sort_syms(num_used_syms, syms0, syms1); + tdefl_calculate_minimum_redundancy(pSyms, num_used_syms); + + for (i = 0; i < num_used_syms; i++) num_codes[pSyms[i].m_key]++; + + tdefl_huffman_enforce_max_code_size(num_codes, num_used_syms, + code_size_limit); + + MZ_CLEAR_OBJ(d->m_huff_code_sizes[table_num]); + MZ_CLEAR_OBJ(d->m_huff_codes[table_num]); + for (i = 1, j = num_used_syms; i <= code_size_limit; i++) + for (l = num_codes[i]; l > 0; l--) + d->m_huff_code_sizes[table_num][pSyms[--j].m_sym_index] = (mz_uint8)(i); + } + + next_code[1] = 0; + for (j = 0, i = 2; i <= code_size_limit; i++) + next_code[i] = j = ((j + num_codes[i - 1]) << 1); + + for (i = 0; i < table_len; i++) { + mz_uint rev_code = 0, code, code_size; + if ((code_size = d->m_huff_code_sizes[table_num][i]) == 0) continue; + code = next_code[code_size]++; + for (l = code_size; l > 0; l--, code >>= 1) + rev_code = (rev_code << 1) | (code & 1); + d->m_huff_codes[table_num][i] = (mz_uint16)rev_code; + } +} + +#define TDEFL_PUT_BITS(b, l) \ + do { \ + mz_uint bits = b; \ + mz_uint len = l; \ + MZ_ASSERT(bits <= ((1U << len) - 1U)); \ + d->m_bit_buffer |= (bits << d->m_bits_in); \ + d->m_bits_in += len; \ + while (d->m_bits_in >= 8) { \ + if (d->m_pOutput_buf < d->m_pOutput_buf_end) \ + *d->m_pOutput_buf++ = (mz_uint8)(d->m_bit_buffer); \ + d->m_bit_buffer >>= 8; \ + d->m_bits_in -= 8; \ + } \ + } \ + MZ_MACRO_END + +#define TDEFL_RLE_PREV_CODE_SIZE() \ + { \ + if (rle_repeat_count) { \ + if (rle_repeat_count < 3) { \ + d->m_huff_count[2][prev_code_size] = (mz_uint16)( \ + d->m_huff_count[2][prev_code_size] + rle_repeat_count); \ + while (rle_repeat_count--) \ + packed_code_sizes[num_packed_code_sizes++] = prev_code_size; \ + } else { \ + d->m_huff_count[2][16] = (mz_uint16)(d->m_huff_count[2][16] + 1); \ + packed_code_sizes[num_packed_code_sizes++] = 16; \ + packed_code_sizes[num_packed_code_sizes++] = \ + (mz_uint8)(rle_repeat_count - 3); \ + } \ + rle_repeat_count = 0; \ + } \ + } + +#define TDEFL_RLE_ZERO_CODE_SIZE() \ + { \ + if (rle_z_count) { \ + if (rle_z_count < 3) { \ + d->m_huff_count[2][0] = \ + (mz_uint16)(d->m_huff_count[2][0] + rle_z_count); \ + while (rle_z_count--) packed_code_sizes[num_packed_code_sizes++] = 0; \ + } else if (rle_z_count <= 10) { \ + d->m_huff_count[2][17] = (mz_uint16)(d->m_huff_count[2][17] + 1); \ + packed_code_sizes[num_packed_code_sizes++] = 17; \ + packed_code_sizes[num_packed_code_sizes++] = \ + (mz_uint8)(rle_z_count - 3); \ + } else { \ + d->m_huff_count[2][18] = (mz_uint16)(d->m_huff_count[2][18] + 1); \ + packed_code_sizes[num_packed_code_sizes++] = 18; \ + packed_code_sizes[num_packed_code_sizes++] = \ + (mz_uint8)(rle_z_count - 11); \ + } \ + rle_z_count = 0; \ + } \ + } + +static mz_uint8 s_tdefl_packed_code_size_syms_swizzle[] = { + 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15}; + +static void tdefl_start_dynamic_block(tdefl_compressor *d) { + int num_lit_codes, num_dist_codes, num_bit_lengths; + mz_uint i, total_code_sizes_to_pack, num_packed_code_sizes, rle_z_count, + rle_repeat_count, packed_code_sizes_index; + mz_uint8 + code_sizes_to_pack[TDEFL_MAX_HUFF_SYMBOLS_0 + TDEFL_MAX_HUFF_SYMBOLS_1], + packed_code_sizes[TDEFL_MAX_HUFF_SYMBOLS_0 + TDEFL_MAX_HUFF_SYMBOLS_1], + prev_code_size = 0xFF; + + d->m_huff_count[0][256] = 1; + + tdefl_optimize_huffman_table(d, 0, TDEFL_MAX_HUFF_SYMBOLS_0, 15, MZ_FALSE); + tdefl_optimize_huffman_table(d, 1, TDEFL_MAX_HUFF_SYMBOLS_1, 15, MZ_FALSE); + + for (num_lit_codes = 286; num_lit_codes > 257; num_lit_codes--) + if (d->m_huff_code_sizes[0][num_lit_codes - 1]) break; + for (num_dist_codes = 30; num_dist_codes > 1; num_dist_codes--) + if (d->m_huff_code_sizes[1][num_dist_codes - 1]) break; + + memcpy(code_sizes_to_pack, &d->m_huff_code_sizes[0][0], num_lit_codes); + memcpy(code_sizes_to_pack + num_lit_codes, &d->m_huff_code_sizes[1][0], + num_dist_codes); + total_code_sizes_to_pack = num_lit_codes + num_dist_codes; + num_packed_code_sizes = 0; + rle_z_count = 0; + rle_repeat_count = 0; + + memset(&d->m_huff_count[2][0], 0, + sizeof(d->m_huff_count[2][0]) * TDEFL_MAX_HUFF_SYMBOLS_2); + for (i = 0; i < total_code_sizes_to_pack; i++) { + mz_uint8 code_size = code_sizes_to_pack[i]; + if (!code_size) { + TDEFL_RLE_PREV_CODE_SIZE(); + if (++rle_z_count == 138) { + TDEFL_RLE_ZERO_CODE_SIZE(); + } + } else { + TDEFL_RLE_ZERO_CODE_SIZE(); + if (code_size != prev_code_size) { + TDEFL_RLE_PREV_CODE_SIZE(); + d->m_huff_count[2][code_size] = + (mz_uint16)(d->m_huff_count[2][code_size] + 1); + packed_code_sizes[num_packed_code_sizes++] = code_size; + } else if (++rle_repeat_count == 6) { + TDEFL_RLE_PREV_CODE_SIZE(); + } + } + prev_code_size = code_size; + } + if (rle_repeat_count) { + TDEFL_RLE_PREV_CODE_SIZE(); + } else { + TDEFL_RLE_ZERO_CODE_SIZE(); + } + + tdefl_optimize_huffman_table(d, 2, TDEFL_MAX_HUFF_SYMBOLS_2, 7, MZ_FALSE); + + TDEFL_PUT_BITS(2, 2); + + TDEFL_PUT_BITS(num_lit_codes - 257, 5); + TDEFL_PUT_BITS(num_dist_codes - 1, 5); + + for (num_bit_lengths = 18; num_bit_lengths >= 0; num_bit_lengths--) + if (d->m_huff_code_sizes + [2][s_tdefl_packed_code_size_syms_swizzle[num_bit_lengths]]) + break; + num_bit_lengths = MZ_MAX(4, (num_bit_lengths + 1)); + TDEFL_PUT_BITS(num_bit_lengths - 4, 4); + for (i = 0; (int)i < num_bit_lengths; i++) + TDEFL_PUT_BITS( + d->m_huff_code_sizes[2][s_tdefl_packed_code_size_syms_swizzle[i]], 3); + + for (packed_code_sizes_index = 0; + packed_code_sizes_index < num_packed_code_sizes;) { + mz_uint code = packed_code_sizes[packed_code_sizes_index++]; + MZ_ASSERT(code < TDEFL_MAX_HUFF_SYMBOLS_2); + TDEFL_PUT_BITS(d->m_huff_codes[2][code], d->m_huff_code_sizes[2][code]); + if (code >= 16) + TDEFL_PUT_BITS(packed_code_sizes[packed_code_sizes_index++], + "\02\03\07"[code - 16]); + } +} + +static void tdefl_start_static_block(tdefl_compressor *d) { + mz_uint i; + mz_uint8 *p = &d->m_huff_code_sizes[0][0]; + + for (i = 0; i <= 143; ++i) *p++ = 8; + for (; i <= 255; ++i) *p++ = 9; + for (; i <= 279; ++i) *p++ = 7; + for (; i <= 287; ++i) *p++ = 8; + + memset(d->m_huff_code_sizes[1], 5, 32); + + tdefl_optimize_huffman_table(d, 0, 288, 15, MZ_TRUE); + tdefl_optimize_huffman_table(d, 1, 32, 15, MZ_TRUE); + + TDEFL_PUT_BITS(1, 2); +} + +static const mz_uint mz_bitmasks[17] = { + 0x0000, 0x0001, 0x0003, 0x0007, 0x000F, 0x001F, 0x003F, 0x007F, 0x00FF, + 0x01FF, 0x03FF, 0x07FF, 0x0FFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF}; + +#if MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN && \ + MINIZ_HAS_64BIT_REGISTERS +static mz_bool tdefl_compress_lz_codes(tdefl_compressor *d) { + mz_uint flags; + mz_uint8 *pLZ_codes; + mz_uint8 *pOutput_buf = d->m_pOutput_buf; + mz_uint8 *pLZ_code_buf_end = d->m_pLZ_code_buf; + mz_uint64 bit_buffer = d->m_bit_buffer; + mz_uint bits_in = d->m_bits_in; + +#define TDEFL_PUT_BITS_FAST(b, l) \ + { \ + bit_buffer |= (((mz_uint64)(b)) << bits_in); \ + bits_in += (l); \ + } + + flags = 1; + for (pLZ_codes = d->m_lz_code_buf; pLZ_codes < pLZ_code_buf_end; + flags >>= 1) { + if (flags == 1) flags = *pLZ_codes++ | 0x100; + + if (flags & 1) { + mz_uint s0, s1, n0, n1, sym, num_extra_bits; + mz_uint match_len = pLZ_codes[0], + match_dist = *(const mz_uint16 *)(pLZ_codes + 1); + pLZ_codes += 3; + + MZ_ASSERT(d->m_huff_code_sizes[0][s_tdefl_len_sym[match_len]]); + TDEFL_PUT_BITS_FAST(d->m_huff_codes[0][s_tdefl_len_sym[match_len]], + d->m_huff_code_sizes[0][s_tdefl_len_sym[match_len]]); + TDEFL_PUT_BITS_FAST(match_len & mz_bitmasks[s_tdefl_len_extra[match_len]], + s_tdefl_len_extra[match_len]); + + // This sequence coaxes MSVC into using cmov's vs. jmp's. + s0 = s_tdefl_small_dist_sym[match_dist & 511]; + n0 = s_tdefl_small_dist_extra[match_dist & 511]; + s1 = s_tdefl_large_dist_sym[match_dist >> 8]; + n1 = s_tdefl_large_dist_extra[match_dist >> 8]; + sym = (match_dist < 512) ? s0 : s1; + num_extra_bits = (match_dist < 512) ? n0 : n1; + + MZ_ASSERT(d->m_huff_code_sizes[1][sym]); + TDEFL_PUT_BITS_FAST(d->m_huff_codes[1][sym], + d->m_huff_code_sizes[1][sym]); + TDEFL_PUT_BITS_FAST(match_dist & mz_bitmasks[num_extra_bits], + num_extra_bits); + } else { + mz_uint lit = *pLZ_codes++; + MZ_ASSERT(d->m_huff_code_sizes[0][lit]); + TDEFL_PUT_BITS_FAST(d->m_huff_codes[0][lit], + d->m_huff_code_sizes[0][lit]); + + if (((flags & 2) == 0) && (pLZ_codes < pLZ_code_buf_end)) { + flags >>= 1; + lit = *pLZ_codes++; + MZ_ASSERT(d->m_huff_code_sizes[0][lit]); + TDEFL_PUT_BITS_FAST(d->m_huff_codes[0][lit], + d->m_huff_code_sizes[0][lit]); + + if (((flags & 2) == 0) && (pLZ_codes < pLZ_code_buf_end)) { + flags >>= 1; + lit = *pLZ_codes++; + MZ_ASSERT(d->m_huff_code_sizes[0][lit]); + TDEFL_PUT_BITS_FAST(d->m_huff_codes[0][lit], + d->m_huff_code_sizes[0][lit]); + } + } + } + + if (pOutput_buf >= d->m_pOutput_buf_end) return MZ_FALSE; + + *(mz_uint64 *)pOutput_buf = bit_buffer; + pOutput_buf += (bits_in >> 3); + bit_buffer >>= (bits_in & ~7); + bits_in &= 7; + } + +#undef TDEFL_PUT_BITS_FAST + + d->m_pOutput_buf = pOutput_buf; + d->m_bits_in = 0; + d->m_bit_buffer = 0; + + while (bits_in) { + mz_uint32 n = MZ_MIN(bits_in, 16); + TDEFL_PUT_BITS((mz_uint)bit_buffer & mz_bitmasks[n], n); + bit_buffer >>= n; + bits_in -= n; + } + + TDEFL_PUT_BITS(d->m_huff_codes[0][256], d->m_huff_code_sizes[0][256]); + + return (d->m_pOutput_buf < d->m_pOutput_buf_end); +} +#else +static mz_bool tdefl_compress_lz_codes(tdefl_compressor *d) { + mz_uint flags; + mz_uint8 *pLZ_codes; + + flags = 1; + for (pLZ_codes = d->m_lz_code_buf; pLZ_codes < d->m_pLZ_code_buf; + flags >>= 1) { + if (flags == 1) flags = *pLZ_codes++ | 0x100; + if (flags & 1) { + mz_uint sym, num_extra_bits; + mz_uint match_len = pLZ_codes[0], + match_dist = (pLZ_codes[1] | (pLZ_codes[2] << 8)); + pLZ_codes += 3; + + MZ_ASSERT(d->m_huff_code_sizes[0][s_tdefl_len_sym[match_len]]); + TDEFL_PUT_BITS(d->m_huff_codes[0][s_tdefl_len_sym[match_len]], + d->m_huff_code_sizes[0][s_tdefl_len_sym[match_len]]); + TDEFL_PUT_BITS(match_len & mz_bitmasks[s_tdefl_len_extra[match_len]], + s_tdefl_len_extra[match_len]); + + if (match_dist < 512) { + sym = s_tdefl_small_dist_sym[match_dist]; + num_extra_bits = s_tdefl_small_dist_extra[match_dist]; + } else { + sym = s_tdefl_large_dist_sym[match_dist >> 8]; + num_extra_bits = s_tdefl_large_dist_extra[match_dist >> 8]; + } + MZ_ASSERT(d->m_huff_code_sizes[1][sym]); + TDEFL_PUT_BITS(d->m_huff_codes[1][sym], d->m_huff_code_sizes[1][sym]); + TDEFL_PUT_BITS(match_dist & mz_bitmasks[num_extra_bits], num_extra_bits); + } else { + mz_uint lit = *pLZ_codes++; + MZ_ASSERT(d->m_huff_code_sizes[0][lit]); + TDEFL_PUT_BITS(d->m_huff_codes[0][lit], d->m_huff_code_sizes[0][lit]); + } + } + + TDEFL_PUT_BITS(d->m_huff_codes[0][256], d->m_huff_code_sizes[0][256]); + + return (d->m_pOutput_buf < d->m_pOutput_buf_end); +} +#endif // MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN && + // MINIZ_HAS_64BIT_REGISTERS + +static mz_bool tdefl_compress_block(tdefl_compressor *d, mz_bool static_block) { + if (static_block) + tdefl_start_static_block(d); + else + tdefl_start_dynamic_block(d); + return tdefl_compress_lz_codes(d); +} + +static int tdefl_flush_block(tdefl_compressor *d, int flush) { + mz_uint saved_bit_buf, saved_bits_in; + mz_uint8 *pSaved_output_buf; + mz_bool comp_block_succeeded = MZ_FALSE; + int n, use_raw_block = + ((d->m_flags & TDEFL_FORCE_ALL_RAW_BLOCKS) != 0) && + (d->m_lookahead_pos - d->m_lz_code_buf_dict_pos) <= d->m_dict_size; + mz_uint8 *pOutput_buf_start = + ((d->m_pPut_buf_func == NULL) && + ((*d->m_pOut_buf_size - d->m_out_buf_ofs) >= TDEFL_OUT_BUF_SIZE)) + ? ((mz_uint8 *)d->m_pOut_buf + d->m_out_buf_ofs) + : d->m_output_buf; + + d->m_pOutput_buf = pOutput_buf_start; + d->m_pOutput_buf_end = d->m_pOutput_buf + TDEFL_OUT_BUF_SIZE - 16; + + MZ_ASSERT(!d->m_output_flush_remaining); + d->m_output_flush_ofs = 0; + d->m_output_flush_remaining = 0; + + *d->m_pLZ_flags = (mz_uint8)(*d->m_pLZ_flags >> d->m_num_flags_left); + d->m_pLZ_code_buf -= (d->m_num_flags_left == 8); + + if ((d->m_flags & TDEFL_WRITE_ZLIB_HEADER) && (!d->m_block_index)) { + TDEFL_PUT_BITS(0x78, 8); + TDEFL_PUT_BITS(0x01, 8); + } + + TDEFL_PUT_BITS(flush == TDEFL_FINISH, 1); + + pSaved_output_buf = d->m_pOutput_buf; + saved_bit_buf = d->m_bit_buffer; + saved_bits_in = d->m_bits_in; + + if (!use_raw_block) + comp_block_succeeded = + tdefl_compress_block(d, (d->m_flags & TDEFL_FORCE_ALL_STATIC_BLOCKS) || + (d->m_total_lz_bytes < 48)); + + // If the block gets expanded, forget the current contents of the output + // buffer and send a raw block instead. + if (((use_raw_block) || + ((d->m_total_lz_bytes) && ((d->m_pOutput_buf - pSaved_output_buf + 1U) >= + d->m_total_lz_bytes))) && + ((d->m_lookahead_pos - d->m_lz_code_buf_dict_pos) <= d->m_dict_size)) { + mz_uint i; + d->m_pOutput_buf = pSaved_output_buf; + d->m_bit_buffer = saved_bit_buf, d->m_bits_in = saved_bits_in; + TDEFL_PUT_BITS(0, 2); + if (d->m_bits_in) { + TDEFL_PUT_BITS(0, 8 - d->m_bits_in); + } + for (i = 2; i; --i, d->m_total_lz_bytes ^= 0xFFFF) { + TDEFL_PUT_BITS(d->m_total_lz_bytes & 0xFFFF, 16); + } + for (i = 0; i < d->m_total_lz_bytes; ++i) { + TDEFL_PUT_BITS( + d->m_dict[(d->m_lz_code_buf_dict_pos + i) & TDEFL_LZ_DICT_SIZE_MASK], + 8); + } + } + // Check for the extremely unlikely (if not impossible) case of the compressed + // block not fitting into the output buffer when using dynamic codes. + else if (!comp_block_succeeded) { + d->m_pOutput_buf = pSaved_output_buf; + d->m_bit_buffer = saved_bit_buf, d->m_bits_in = saved_bits_in; + tdefl_compress_block(d, MZ_TRUE); + } + + if (flush) { + if (flush == TDEFL_FINISH) { + if (d->m_bits_in) { + TDEFL_PUT_BITS(0, 8 - d->m_bits_in); + } + if (d->m_flags & TDEFL_WRITE_ZLIB_HEADER) { + mz_uint i, a = d->m_adler32; + for (i = 0; i < 4; i++) { + TDEFL_PUT_BITS((a >> 24) & 0xFF, 8); + a <<= 8; + } + } + } else { + mz_uint i, z = 0; + TDEFL_PUT_BITS(0, 3); + if (d->m_bits_in) { + TDEFL_PUT_BITS(0, 8 - d->m_bits_in); + } + for (i = 2; i; --i, z ^= 0xFFFF) { + TDEFL_PUT_BITS(z & 0xFFFF, 16); + } + } + } + + MZ_ASSERT(d->m_pOutput_buf < d->m_pOutput_buf_end); + + memset(&d->m_huff_count[0][0], 0, + sizeof(d->m_huff_count[0][0]) * TDEFL_MAX_HUFF_SYMBOLS_0); + memset(&d->m_huff_count[1][0], 0, + sizeof(d->m_huff_count[1][0]) * TDEFL_MAX_HUFF_SYMBOLS_1); + + d->m_pLZ_code_buf = d->m_lz_code_buf + 1; + d->m_pLZ_flags = d->m_lz_code_buf; + d->m_num_flags_left = 8; + d->m_lz_code_buf_dict_pos += d->m_total_lz_bytes; + d->m_total_lz_bytes = 0; + d->m_block_index++; + + if ((n = (int)(d->m_pOutput_buf - pOutput_buf_start)) != 0) { + if (d->m_pPut_buf_func) { + *d->m_pIn_buf_size = d->m_pSrc - (const mz_uint8 *)d->m_pIn_buf; + if (!(*d->m_pPut_buf_func)(d->m_output_buf, n, d->m_pPut_buf_user)) + return (d->m_prev_return_status = TDEFL_STATUS_PUT_BUF_FAILED); + } else if (pOutput_buf_start == d->m_output_buf) { + int bytes_to_copy = (int)MZ_MIN( + (size_t)n, (size_t)(*d->m_pOut_buf_size - d->m_out_buf_ofs)); + memcpy((mz_uint8 *)d->m_pOut_buf + d->m_out_buf_ofs, d->m_output_buf, + bytes_to_copy); + d->m_out_buf_ofs += bytes_to_copy; + if ((n -= bytes_to_copy) != 0) { + d->m_output_flush_ofs = bytes_to_copy; + d->m_output_flush_remaining = n; + } + } else { + d->m_out_buf_ofs += n; + } + } + + return d->m_output_flush_remaining; +} + +#if MINIZ_USE_UNALIGNED_LOADS_AND_STORES +#define TDEFL_READ_UNALIGNED_WORD(p) *(const mz_uint16 *)(p) +static MZ_FORCEINLINE void tdefl_find_match( + tdefl_compressor *d, mz_uint lookahead_pos, mz_uint max_dist, + mz_uint max_match_len, mz_uint *pMatch_dist, mz_uint *pMatch_len) { + mz_uint dist, pos = lookahead_pos & TDEFL_LZ_DICT_SIZE_MASK, + match_len = *pMatch_len, probe_pos = pos, next_probe_pos, + probe_len; + mz_uint num_probes_left = d->m_max_probes[match_len >= 32]; + const mz_uint16 *s = (const mz_uint16 *)(d->m_dict + pos), *p, *q; + mz_uint16 c01 = TDEFL_READ_UNALIGNED_WORD(&d->m_dict[pos + match_len - 1]), + s01 = TDEFL_READ_UNALIGNED_WORD(s); + MZ_ASSERT(max_match_len <= TDEFL_MAX_MATCH_LEN); + if (max_match_len <= match_len) return; + for (;;) { + for (;;) { + if (--num_probes_left == 0) return; +#define TDEFL_PROBE \ + next_probe_pos = d->m_next[probe_pos]; \ + if ((!next_probe_pos) || \ + ((dist = (mz_uint16)(lookahead_pos - next_probe_pos)) > max_dist)) \ + return; \ + probe_pos = next_probe_pos & TDEFL_LZ_DICT_SIZE_MASK; \ + if (TDEFL_READ_UNALIGNED_WORD(&d->m_dict[probe_pos + match_len - 1]) == c01) \ + break; + TDEFL_PROBE; + TDEFL_PROBE; + TDEFL_PROBE; + } + if (!dist) break; + q = (const mz_uint16 *)(d->m_dict + probe_pos); + if (TDEFL_READ_UNALIGNED_WORD(q) != s01) continue; + p = s; + probe_len = 32; + do { + } while ( + (TDEFL_READ_UNALIGNED_WORD(++p) == TDEFL_READ_UNALIGNED_WORD(++q)) && + (TDEFL_READ_UNALIGNED_WORD(++p) == TDEFL_READ_UNALIGNED_WORD(++q)) && + (TDEFL_READ_UNALIGNED_WORD(++p) == TDEFL_READ_UNALIGNED_WORD(++q)) && + (TDEFL_READ_UNALIGNED_WORD(++p) == TDEFL_READ_UNALIGNED_WORD(++q)) && + (--probe_len > 0)); + if (!probe_len) { + *pMatch_dist = dist; + *pMatch_len = MZ_MIN(max_match_len, TDEFL_MAX_MATCH_LEN); + break; + } else if ((probe_len = ((mz_uint)(p - s) * 2) + + (mz_uint)(*(const mz_uint8 *)p == + *(const mz_uint8 *)q)) > match_len) { + *pMatch_dist = dist; + if ((*pMatch_len = match_len = MZ_MIN(max_match_len, probe_len)) == + max_match_len) + break; + c01 = TDEFL_READ_UNALIGNED_WORD(&d->m_dict[pos + match_len - 1]); + } + } +} +#else +static MZ_FORCEINLINE void tdefl_find_match( + tdefl_compressor *d, mz_uint lookahead_pos, mz_uint max_dist, + mz_uint max_match_len, mz_uint *pMatch_dist, mz_uint *pMatch_len) { + mz_uint dist, pos = lookahead_pos & TDEFL_LZ_DICT_SIZE_MASK, + match_len = *pMatch_len, probe_pos = pos, next_probe_pos, + probe_len; + mz_uint num_probes_left = d->m_max_probes[match_len >= 32]; + const mz_uint8 *s = d->m_dict + pos, *p, *q; + mz_uint8 c0 = d->m_dict[pos + match_len], c1 = d->m_dict[pos + match_len - 1]; + MZ_ASSERT(max_match_len <= TDEFL_MAX_MATCH_LEN); + if (max_match_len <= match_len) return; + for (;;) { + for (;;) { + if (--num_probes_left == 0) return; +#define TDEFL_PROBE \ + next_probe_pos = d->m_next[probe_pos]; \ + if ((!next_probe_pos) || \ + ((dist = (mz_uint16)(lookahead_pos - next_probe_pos)) > max_dist)) \ + return; \ + probe_pos = next_probe_pos & TDEFL_LZ_DICT_SIZE_MASK; \ + if ((d->m_dict[probe_pos + match_len] == c0) && \ + (d->m_dict[probe_pos + match_len - 1] == c1)) \ + break; + TDEFL_PROBE; + TDEFL_PROBE; + TDEFL_PROBE; + } + if (!dist) break; + p = s; + q = d->m_dict + probe_pos; + for (probe_len = 0; probe_len < max_match_len; probe_len++) + if (*p++ != *q++) break; + if (probe_len > match_len) { + *pMatch_dist = dist; + if ((*pMatch_len = match_len = probe_len) == max_match_len) return; + c0 = d->m_dict[pos + match_len]; + c1 = d->m_dict[pos + match_len - 1]; + } + } +} +#endif // #if MINIZ_USE_UNALIGNED_LOADS_AND_STORES + +#if MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN +static mz_bool tdefl_compress_fast(tdefl_compressor *d) { + // Faster, minimally featured LZRW1-style match+parse loop with better + // register utilization. Intended for applications where raw throughput is + // valued more highly than ratio. + mz_uint lookahead_pos = d->m_lookahead_pos, + lookahead_size = d->m_lookahead_size, dict_size = d->m_dict_size, + total_lz_bytes = d->m_total_lz_bytes, + num_flags_left = d->m_num_flags_left; + mz_uint8 *pLZ_code_buf = d->m_pLZ_code_buf, *pLZ_flags = d->m_pLZ_flags; + mz_uint cur_pos = lookahead_pos & TDEFL_LZ_DICT_SIZE_MASK; + + while ((d->m_src_buf_left) || ((d->m_flush) && (lookahead_size))) { + const mz_uint TDEFL_COMP_FAST_LOOKAHEAD_SIZE = 4096; + mz_uint dst_pos = + (lookahead_pos + lookahead_size) & TDEFL_LZ_DICT_SIZE_MASK; + mz_uint num_bytes_to_process = (mz_uint)MZ_MIN( + d->m_src_buf_left, TDEFL_COMP_FAST_LOOKAHEAD_SIZE - lookahead_size); + d->m_src_buf_left -= num_bytes_to_process; + lookahead_size += num_bytes_to_process; + + while (num_bytes_to_process) { + mz_uint32 n = MZ_MIN(TDEFL_LZ_DICT_SIZE - dst_pos, num_bytes_to_process); + memcpy(d->m_dict + dst_pos, d->m_pSrc, n); + if (dst_pos < (TDEFL_MAX_MATCH_LEN - 1)) + memcpy(d->m_dict + TDEFL_LZ_DICT_SIZE + dst_pos, d->m_pSrc, + MZ_MIN(n, (TDEFL_MAX_MATCH_LEN - 1) - dst_pos)); + d->m_pSrc += n; + dst_pos = (dst_pos + n) & TDEFL_LZ_DICT_SIZE_MASK; + num_bytes_to_process -= n; + } + + dict_size = MZ_MIN(TDEFL_LZ_DICT_SIZE - lookahead_size, dict_size); + if ((!d->m_flush) && (lookahead_size < TDEFL_COMP_FAST_LOOKAHEAD_SIZE)) + break; + + while (lookahead_size >= 4) { + mz_uint cur_match_dist, cur_match_len = 1; + mz_uint8 *pCur_dict = d->m_dict + cur_pos; + mz_uint first_trigram = (*(const mz_uint32 *)pCur_dict) & 0xFFFFFF; + mz_uint hash = + (first_trigram ^ (first_trigram >> (24 - (TDEFL_LZ_HASH_BITS - 8)))) & + TDEFL_LEVEL1_HASH_SIZE_MASK; + mz_uint probe_pos = d->m_hash[hash]; + d->m_hash[hash] = (mz_uint16)lookahead_pos; + + if (((cur_match_dist = (mz_uint16)(lookahead_pos - probe_pos)) <= + dict_size) && + ((*(const mz_uint32 *)(d->m_dict + + (probe_pos &= TDEFL_LZ_DICT_SIZE_MASK)) & + 0xFFFFFF) == first_trigram)) { + const mz_uint16 *p = (const mz_uint16 *)pCur_dict; + const mz_uint16 *q = (const mz_uint16 *)(d->m_dict + probe_pos); + mz_uint32 probe_len = 32; + do { + } while ((TDEFL_READ_UNALIGNED_WORD(++p) == + TDEFL_READ_UNALIGNED_WORD(++q)) && + (TDEFL_READ_UNALIGNED_WORD(++p) == + TDEFL_READ_UNALIGNED_WORD(++q)) && + (TDEFL_READ_UNALIGNED_WORD(++p) == + TDEFL_READ_UNALIGNED_WORD(++q)) && + (TDEFL_READ_UNALIGNED_WORD(++p) == + TDEFL_READ_UNALIGNED_WORD(++q)) && + (--probe_len > 0)); + cur_match_len = ((mz_uint)(p - (const mz_uint16 *)pCur_dict) * 2) + + (mz_uint)(*(const mz_uint8 *)p == *(const mz_uint8 *)q); + if (!probe_len) + cur_match_len = cur_match_dist ? TDEFL_MAX_MATCH_LEN : 0; + + if ((cur_match_len < TDEFL_MIN_MATCH_LEN) || + ((cur_match_len == TDEFL_MIN_MATCH_LEN) && + (cur_match_dist >= 8U * 1024U))) { + cur_match_len = 1; + *pLZ_code_buf++ = (mz_uint8)first_trigram; + *pLZ_flags = (mz_uint8)(*pLZ_flags >> 1); + d->m_huff_count[0][(mz_uint8)first_trigram]++; + } else { + mz_uint32 s0, s1; + cur_match_len = MZ_MIN(cur_match_len, lookahead_size); + + MZ_ASSERT((cur_match_len >= TDEFL_MIN_MATCH_LEN) && + (cur_match_dist >= 1) && + (cur_match_dist <= TDEFL_LZ_DICT_SIZE)); + + cur_match_dist--; + + pLZ_code_buf[0] = (mz_uint8)(cur_match_len - TDEFL_MIN_MATCH_LEN); + *(mz_uint16 *)(&pLZ_code_buf[1]) = (mz_uint16)cur_match_dist; + pLZ_code_buf += 3; + *pLZ_flags = (mz_uint8)((*pLZ_flags >> 1) | 0x80); + + s0 = s_tdefl_small_dist_sym[cur_match_dist & 511]; + s1 = s_tdefl_large_dist_sym[cur_match_dist >> 8]; + d->m_huff_count[1][(cur_match_dist < 512) ? s0 : s1]++; + + d->m_huff_count[0][s_tdefl_len_sym[cur_match_len - + TDEFL_MIN_MATCH_LEN]]++; + } + } else { + *pLZ_code_buf++ = (mz_uint8)first_trigram; + *pLZ_flags = (mz_uint8)(*pLZ_flags >> 1); + d->m_huff_count[0][(mz_uint8)first_trigram]++; + } + + if (--num_flags_left == 0) { + num_flags_left = 8; + pLZ_flags = pLZ_code_buf++; + } + + total_lz_bytes += cur_match_len; + lookahead_pos += cur_match_len; + dict_size = MZ_MIN(dict_size + cur_match_len, TDEFL_LZ_DICT_SIZE); + cur_pos = (cur_pos + cur_match_len) & TDEFL_LZ_DICT_SIZE_MASK; + MZ_ASSERT(lookahead_size >= cur_match_len); + lookahead_size -= cur_match_len; + + if (pLZ_code_buf > &d->m_lz_code_buf[TDEFL_LZ_CODE_BUF_SIZE - 8]) { + int n; + d->m_lookahead_pos = lookahead_pos; + d->m_lookahead_size = lookahead_size; + d->m_dict_size = dict_size; + d->m_total_lz_bytes = total_lz_bytes; + d->m_pLZ_code_buf = pLZ_code_buf; + d->m_pLZ_flags = pLZ_flags; + d->m_num_flags_left = num_flags_left; + if ((n = tdefl_flush_block(d, 0)) != 0) + return (n < 0) ? MZ_FALSE : MZ_TRUE; + total_lz_bytes = d->m_total_lz_bytes; + pLZ_code_buf = d->m_pLZ_code_buf; + pLZ_flags = d->m_pLZ_flags; + num_flags_left = d->m_num_flags_left; + } + } + + while (lookahead_size) { + mz_uint8 lit = d->m_dict[cur_pos]; + + total_lz_bytes++; + *pLZ_code_buf++ = lit; + *pLZ_flags = (mz_uint8)(*pLZ_flags >> 1); + if (--num_flags_left == 0) { + num_flags_left = 8; + pLZ_flags = pLZ_code_buf++; + } + + d->m_huff_count[0][lit]++; + + lookahead_pos++; + dict_size = MZ_MIN(dict_size + 1, TDEFL_LZ_DICT_SIZE); + cur_pos = (cur_pos + 1) & TDEFL_LZ_DICT_SIZE_MASK; + lookahead_size--; + + if (pLZ_code_buf > &d->m_lz_code_buf[TDEFL_LZ_CODE_BUF_SIZE - 8]) { + int n; + d->m_lookahead_pos = lookahead_pos; + d->m_lookahead_size = lookahead_size; + d->m_dict_size = dict_size; + d->m_total_lz_bytes = total_lz_bytes; + d->m_pLZ_code_buf = pLZ_code_buf; + d->m_pLZ_flags = pLZ_flags; + d->m_num_flags_left = num_flags_left; + if ((n = tdefl_flush_block(d, 0)) != 0) + return (n < 0) ? MZ_FALSE : MZ_TRUE; + total_lz_bytes = d->m_total_lz_bytes; + pLZ_code_buf = d->m_pLZ_code_buf; + pLZ_flags = d->m_pLZ_flags; + num_flags_left = d->m_num_flags_left; + } + } + } + + d->m_lookahead_pos = lookahead_pos; + d->m_lookahead_size = lookahead_size; + d->m_dict_size = dict_size; + d->m_total_lz_bytes = total_lz_bytes; + d->m_pLZ_code_buf = pLZ_code_buf; + d->m_pLZ_flags = pLZ_flags; + d->m_num_flags_left = num_flags_left; + return MZ_TRUE; +} +#endif // MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN + +static MZ_FORCEINLINE void tdefl_record_literal(tdefl_compressor *d, + mz_uint8 lit) { + d->m_total_lz_bytes++; + *d->m_pLZ_code_buf++ = lit; + *d->m_pLZ_flags = (mz_uint8)(*d->m_pLZ_flags >> 1); + if (--d->m_num_flags_left == 0) { + d->m_num_flags_left = 8; + d->m_pLZ_flags = d->m_pLZ_code_buf++; + } + d->m_huff_count[0][lit]++; +} + +static MZ_FORCEINLINE void tdefl_record_match(tdefl_compressor *d, + mz_uint match_len, + mz_uint match_dist) { + mz_uint32 s0, s1; + + MZ_ASSERT((match_len >= TDEFL_MIN_MATCH_LEN) && (match_dist >= 1) && + (match_dist <= TDEFL_LZ_DICT_SIZE)); + + d->m_total_lz_bytes += match_len; + + d->m_pLZ_code_buf[0] = (mz_uint8)(match_len - TDEFL_MIN_MATCH_LEN); + + match_dist -= 1; + d->m_pLZ_code_buf[1] = (mz_uint8)(match_dist & 0xFF); + d->m_pLZ_code_buf[2] = (mz_uint8)(match_dist >> 8); + d->m_pLZ_code_buf += 3; + + *d->m_pLZ_flags = (mz_uint8)((*d->m_pLZ_flags >> 1) | 0x80); + if (--d->m_num_flags_left == 0) { + d->m_num_flags_left = 8; + d->m_pLZ_flags = d->m_pLZ_code_buf++; + } + + s0 = s_tdefl_small_dist_sym[match_dist & 511]; + s1 = s_tdefl_large_dist_sym[(match_dist >> 8) & 127]; + d->m_huff_count[1][(match_dist < 512) ? s0 : s1]++; + + if (match_len >= TDEFL_MIN_MATCH_LEN) + d->m_huff_count[0][s_tdefl_len_sym[match_len - TDEFL_MIN_MATCH_LEN]]++; +} + +static mz_bool tdefl_compress_normal(tdefl_compressor *d) { + const mz_uint8 *pSrc = d->m_pSrc; + size_t src_buf_left = d->m_src_buf_left; + tdefl_flush flush = d->m_flush; + + while ((src_buf_left) || ((flush) && (d->m_lookahead_size))) { + mz_uint len_to_move, cur_match_dist, cur_match_len, cur_pos; + // Update dictionary and hash chains. Keeps the lookahead size equal to + // TDEFL_MAX_MATCH_LEN. + if ((d->m_lookahead_size + d->m_dict_size) >= (TDEFL_MIN_MATCH_LEN - 1)) { + mz_uint dst_pos = (d->m_lookahead_pos + d->m_lookahead_size) & + TDEFL_LZ_DICT_SIZE_MASK, + ins_pos = d->m_lookahead_pos + d->m_lookahead_size - 2; + mz_uint hash = (d->m_dict[ins_pos & TDEFL_LZ_DICT_SIZE_MASK] + << TDEFL_LZ_HASH_SHIFT) ^ + d->m_dict[(ins_pos + 1) & TDEFL_LZ_DICT_SIZE_MASK]; + mz_uint num_bytes_to_process = (mz_uint)MZ_MIN( + src_buf_left, TDEFL_MAX_MATCH_LEN - d->m_lookahead_size); + const mz_uint8 *pSrc_end = pSrc + num_bytes_to_process; + src_buf_left -= num_bytes_to_process; + d->m_lookahead_size += num_bytes_to_process; + while (pSrc != pSrc_end) { + mz_uint8 c = *pSrc++; + d->m_dict[dst_pos] = c; + if (dst_pos < (TDEFL_MAX_MATCH_LEN - 1)) + d->m_dict[TDEFL_LZ_DICT_SIZE + dst_pos] = c; + hash = ((hash << TDEFL_LZ_HASH_SHIFT) ^ c) & (TDEFL_LZ_HASH_SIZE - 1); + d->m_next[ins_pos & TDEFL_LZ_DICT_SIZE_MASK] = d->m_hash[hash]; + d->m_hash[hash] = (mz_uint16)(ins_pos); + dst_pos = (dst_pos + 1) & TDEFL_LZ_DICT_SIZE_MASK; + ins_pos++; + } + } else { + while ((src_buf_left) && (d->m_lookahead_size < TDEFL_MAX_MATCH_LEN)) { + mz_uint8 c = *pSrc++; + mz_uint dst_pos = (d->m_lookahead_pos + d->m_lookahead_size) & + TDEFL_LZ_DICT_SIZE_MASK; + src_buf_left--; + d->m_dict[dst_pos] = c; + if (dst_pos < (TDEFL_MAX_MATCH_LEN - 1)) + d->m_dict[TDEFL_LZ_DICT_SIZE + dst_pos] = c; + if ((++d->m_lookahead_size + d->m_dict_size) >= TDEFL_MIN_MATCH_LEN) { + mz_uint ins_pos = d->m_lookahead_pos + (d->m_lookahead_size - 1) - 2; + mz_uint hash = ((d->m_dict[ins_pos & TDEFL_LZ_DICT_SIZE_MASK] + << (TDEFL_LZ_HASH_SHIFT * 2)) ^ + (d->m_dict[(ins_pos + 1) & TDEFL_LZ_DICT_SIZE_MASK] + << TDEFL_LZ_HASH_SHIFT) ^ + c) & + (TDEFL_LZ_HASH_SIZE - 1); + d->m_next[ins_pos & TDEFL_LZ_DICT_SIZE_MASK] = d->m_hash[hash]; + d->m_hash[hash] = (mz_uint16)(ins_pos); + } + } + } + d->m_dict_size = + MZ_MIN(TDEFL_LZ_DICT_SIZE - d->m_lookahead_size, d->m_dict_size); + if ((!flush) && (d->m_lookahead_size < TDEFL_MAX_MATCH_LEN)) break; + + // Simple lazy/greedy parsing state machine. + len_to_move = 1; + cur_match_dist = 0; + cur_match_len = + d->m_saved_match_len ? d->m_saved_match_len : (TDEFL_MIN_MATCH_LEN - 1); + cur_pos = d->m_lookahead_pos & TDEFL_LZ_DICT_SIZE_MASK; + if (d->m_flags & (TDEFL_RLE_MATCHES | TDEFL_FORCE_ALL_RAW_BLOCKS)) { + if ((d->m_dict_size) && (!(d->m_flags & TDEFL_FORCE_ALL_RAW_BLOCKS))) { + mz_uint8 c = d->m_dict[(cur_pos - 1) & TDEFL_LZ_DICT_SIZE_MASK]; + cur_match_len = 0; + while (cur_match_len < d->m_lookahead_size) { + if (d->m_dict[cur_pos + cur_match_len] != c) break; + cur_match_len++; + } + if (cur_match_len < TDEFL_MIN_MATCH_LEN) + cur_match_len = 0; + else + cur_match_dist = 1; + } + } else { + tdefl_find_match(d, d->m_lookahead_pos, d->m_dict_size, + d->m_lookahead_size, &cur_match_dist, &cur_match_len); + } + if (((cur_match_len == TDEFL_MIN_MATCH_LEN) && + (cur_match_dist >= 8U * 1024U)) || + (cur_pos == cur_match_dist) || + ((d->m_flags & TDEFL_FILTER_MATCHES) && (cur_match_len <= 5))) { + cur_match_dist = cur_match_len = 0; + } + if (d->m_saved_match_len) { + if (cur_match_len > d->m_saved_match_len) { + tdefl_record_literal(d, (mz_uint8)d->m_saved_lit); + if (cur_match_len >= 128) { + tdefl_record_match(d, cur_match_len, cur_match_dist); + d->m_saved_match_len = 0; + len_to_move = cur_match_len; + } else { + d->m_saved_lit = d->m_dict[cur_pos]; + d->m_saved_match_dist = cur_match_dist; + d->m_saved_match_len = cur_match_len; + } + } else { + tdefl_record_match(d, d->m_saved_match_len, d->m_saved_match_dist); + len_to_move = d->m_saved_match_len - 1; + d->m_saved_match_len = 0; + } + } else if (!cur_match_dist) + tdefl_record_literal(d, + d->m_dict[MZ_MIN(cur_pos, sizeof(d->m_dict) - 1)]); + else if ((d->m_greedy_parsing) || (d->m_flags & TDEFL_RLE_MATCHES) || + (cur_match_len >= 128)) { + tdefl_record_match(d, cur_match_len, cur_match_dist); + len_to_move = cur_match_len; + } else { + d->m_saved_lit = d->m_dict[MZ_MIN(cur_pos, sizeof(d->m_dict) - 1)]; + d->m_saved_match_dist = cur_match_dist; + d->m_saved_match_len = cur_match_len; + } + // Move the lookahead forward by len_to_move bytes. + d->m_lookahead_pos += len_to_move; + MZ_ASSERT(d->m_lookahead_size >= len_to_move); + d->m_lookahead_size -= len_to_move; + d->m_dict_size = + MZ_MIN(d->m_dict_size + len_to_move, (mz_uint)TDEFL_LZ_DICT_SIZE); + // Check if it's time to flush the current LZ codes to the internal output + // buffer. + if ((d->m_pLZ_code_buf > &d->m_lz_code_buf[TDEFL_LZ_CODE_BUF_SIZE - 8]) || + ((d->m_total_lz_bytes > 31 * 1024) && + (((((mz_uint)(d->m_pLZ_code_buf - d->m_lz_code_buf) * 115) >> 7) >= + d->m_total_lz_bytes) || + (d->m_flags & TDEFL_FORCE_ALL_RAW_BLOCKS)))) { + int n; + d->m_pSrc = pSrc; + d->m_src_buf_left = src_buf_left; + if ((n = tdefl_flush_block(d, 0)) != 0) + return (n < 0) ? MZ_FALSE : MZ_TRUE; + } + } + + d->m_pSrc = pSrc; + d->m_src_buf_left = src_buf_left; + return MZ_TRUE; +} + +static tdefl_status tdefl_flush_output_buffer(tdefl_compressor *d) { + if (d->m_pIn_buf_size) { + *d->m_pIn_buf_size = d->m_pSrc - (const mz_uint8 *)d->m_pIn_buf; + } + + if (d->m_pOut_buf_size) { + size_t n = MZ_MIN(*d->m_pOut_buf_size - d->m_out_buf_ofs, + d->m_output_flush_remaining); + memcpy((mz_uint8 *)d->m_pOut_buf + d->m_out_buf_ofs, + d->m_output_buf + d->m_output_flush_ofs, n); + d->m_output_flush_ofs += (mz_uint)n; + d->m_output_flush_remaining -= (mz_uint)n; + d->m_out_buf_ofs += n; + + *d->m_pOut_buf_size = d->m_out_buf_ofs; + } + + return (d->m_finished && !d->m_output_flush_remaining) ? TDEFL_STATUS_DONE + : TDEFL_STATUS_OKAY; +} + +tdefl_status tdefl_compress(tdefl_compressor *d, const void *pIn_buf, + size_t *pIn_buf_size, void *pOut_buf, + size_t *pOut_buf_size, tdefl_flush flush) { + if (!d) { + if (pIn_buf_size) *pIn_buf_size = 0; + if (pOut_buf_size) *pOut_buf_size = 0; + return TDEFL_STATUS_BAD_PARAM; + } + + d->m_pIn_buf = pIn_buf; + d->m_pIn_buf_size = pIn_buf_size; + d->m_pOut_buf = pOut_buf; + d->m_pOut_buf_size = pOut_buf_size; + d->m_pSrc = (const mz_uint8 *)(pIn_buf); + d->m_src_buf_left = pIn_buf_size ? *pIn_buf_size : 0; + d->m_out_buf_ofs = 0; + d->m_flush = flush; + + if (((d->m_pPut_buf_func != NULL) == + ((pOut_buf != NULL) || (pOut_buf_size != NULL))) || + (d->m_prev_return_status != TDEFL_STATUS_OKAY) || + (d->m_wants_to_finish && (flush != TDEFL_FINISH)) || + (pIn_buf_size && *pIn_buf_size && !pIn_buf) || + (pOut_buf_size && *pOut_buf_size && !pOut_buf)) { + if (pIn_buf_size) *pIn_buf_size = 0; + if (pOut_buf_size) *pOut_buf_size = 0; + return (d->m_prev_return_status = TDEFL_STATUS_BAD_PARAM); + } + d->m_wants_to_finish |= (flush == TDEFL_FINISH); + + if ((d->m_output_flush_remaining) || (d->m_finished)) + return (d->m_prev_return_status = tdefl_flush_output_buffer(d)); + +#if MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN + if (((d->m_flags & TDEFL_MAX_PROBES_MASK) == 1) && + ((d->m_flags & TDEFL_GREEDY_PARSING_FLAG) != 0) && + ((d->m_flags & (TDEFL_FILTER_MATCHES | TDEFL_FORCE_ALL_RAW_BLOCKS | + TDEFL_RLE_MATCHES)) == 0)) { + if (!tdefl_compress_fast(d)) return d->m_prev_return_status; + } else +#endif // #if MINIZ_USE_UNALIGNED_LOADS_AND_STORES && MINIZ_LITTLE_ENDIAN + { + if (!tdefl_compress_normal(d)) return d->m_prev_return_status; + } + + if ((d->m_flags & (TDEFL_WRITE_ZLIB_HEADER | TDEFL_COMPUTE_ADLER32)) && + (pIn_buf)) + d->m_adler32 = + (mz_uint32)mz_adler32(d->m_adler32, (const mz_uint8 *)pIn_buf, + d->m_pSrc - (const mz_uint8 *)pIn_buf); + + if ((flush) && (!d->m_lookahead_size) && (!d->m_src_buf_left) && + (!d->m_output_flush_remaining)) { + if (tdefl_flush_block(d, flush) < 0) return d->m_prev_return_status; + d->m_finished = (flush == TDEFL_FINISH); + if (flush == TDEFL_FULL_FLUSH) { + MZ_CLEAR_OBJ(d->m_hash); + MZ_CLEAR_OBJ(d->m_next); + d->m_dict_size = 0; + } + } + + return (d->m_prev_return_status = tdefl_flush_output_buffer(d)); +} + +tdefl_status tdefl_compress_buffer(tdefl_compressor *d, const void *pIn_buf, + size_t in_buf_size, tdefl_flush flush) { + MZ_ASSERT(d->m_pPut_buf_func); + return tdefl_compress(d, pIn_buf, &in_buf_size, NULL, NULL, flush); +} + +tdefl_status tdefl_init(tdefl_compressor *d, + tdefl_put_buf_func_ptr pPut_buf_func, + void *pPut_buf_user, int flags) { + d->m_pPut_buf_func = pPut_buf_func; + d->m_pPut_buf_user = pPut_buf_user; + d->m_flags = (mz_uint)(flags); + d->m_max_probes[0] = 1 + ((flags & 0xFFF) + 2) / 3; + d->m_greedy_parsing = (flags & TDEFL_GREEDY_PARSING_FLAG) != 0; + d->m_max_probes[1] = 1 + (((flags & 0xFFF) >> 2) + 2) / 3; + if (!(flags & TDEFL_NONDETERMINISTIC_PARSING_FLAG)) MZ_CLEAR_OBJ(d->m_hash); + d->m_lookahead_pos = d->m_lookahead_size = d->m_dict_size = + d->m_total_lz_bytes = d->m_lz_code_buf_dict_pos = d->m_bits_in = 0; + d->m_output_flush_ofs = d->m_output_flush_remaining = d->m_finished = + d->m_block_index = d->m_bit_buffer = d->m_wants_to_finish = 0; + d->m_pLZ_code_buf = d->m_lz_code_buf + 1; + d->m_pLZ_flags = d->m_lz_code_buf; + d->m_num_flags_left = 8; + d->m_pOutput_buf = d->m_output_buf; + d->m_pOutput_buf_end = d->m_output_buf; + d->m_prev_return_status = TDEFL_STATUS_OKAY; + d->m_saved_match_dist = d->m_saved_match_len = d->m_saved_lit = 0; + d->m_adler32 = 1; + d->m_pIn_buf = NULL; + d->m_pOut_buf = NULL; + d->m_pIn_buf_size = NULL; + d->m_pOut_buf_size = NULL; + d->m_flush = TDEFL_NO_FLUSH; + d->m_pSrc = NULL; + d->m_src_buf_left = 0; + d->m_out_buf_ofs = 0; + memset(&d->m_huff_count[0][0], 0, + sizeof(d->m_huff_count[0][0]) * TDEFL_MAX_HUFF_SYMBOLS_0); + memset(&d->m_huff_count[1][0], 0, + sizeof(d->m_huff_count[1][0]) * TDEFL_MAX_HUFF_SYMBOLS_1); + return TDEFL_STATUS_OKAY; +} + +tdefl_status tdefl_get_prev_return_status(tdefl_compressor *d) { + return d->m_prev_return_status; +} + +mz_uint32 tdefl_get_adler32(tdefl_compressor *d) { return d->m_adler32; } + +mz_bool tdefl_compress_mem_to_output(const void *pBuf, size_t buf_len, + tdefl_put_buf_func_ptr pPut_buf_func, + void *pPut_buf_user, int flags) { + tdefl_compressor *pComp; + mz_bool succeeded; + if (((buf_len) && (!pBuf)) || (!pPut_buf_func)) return MZ_FALSE; + pComp = (tdefl_compressor *)MZ_MALLOC(sizeof(tdefl_compressor)); + if (!pComp) return MZ_FALSE; + succeeded = (tdefl_init(pComp, pPut_buf_func, pPut_buf_user, flags) == + TDEFL_STATUS_OKAY); + succeeded = + succeeded && (tdefl_compress_buffer(pComp, pBuf, buf_len, TDEFL_FINISH) == + TDEFL_STATUS_DONE); + MZ_FREE(pComp); + return succeeded; +} + +typedef struct { + size_t m_size, m_capacity; + mz_uint8 *m_pBuf; + mz_bool m_expandable; +} tdefl_output_buffer; + +static mz_bool tdefl_output_buffer_putter(const void *pBuf, int len, + void *pUser) { + tdefl_output_buffer *p = (tdefl_output_buffer *)pUser; + size_t new_size = p->m_size + len; + if (new_size > p->m_capacity) { + size_t new_capacity = p->m_capacity; + mz_uint8 *pNew_buf; + if (!p->m_expandable) return MZ_FALSE; + do { + new_capacity = MZ_MAX(128U, new_capacity << 1U); + } while (new_size > new_capacity); + pNew_buf = (mz_uint8 *)MZ_REALLOC(p->m_pBuf, new_capacity); + if (!pNew_buf) return MZ_FALSE; + p->m_pBuf = pNew_buf; + p->m_capacity = new_capacity; + } + memcpy((mz_uint8 *)p->m_pBuf + p->m_size, pBuf, len); + p->m_size = new_size; + return MZ_TRUE; +} + +void *tdefl_compress_mem_to_heap(const void *pSrc_buf, size_t src_buf_len, + size_t *pOut_len, int flags) { + tdefl_output_buffer out_buf; + MZ_CLEAR_OBJ(out_buf); + if (!pOut_len) + return MZ_FALSE; + else + *pOut_len = 0; + out_buf.m_expandable = MZ_TRUE; + if (!tdefl_compress_mem_to_output( + pSrc_buf, src_buf_len, tdefl_output_buffer_putter, &out_buf, flags)) + return NULL; + *pOut_len = out_buf.m_size; + return out_buf.m_pBuf; +} + +size_t tdefl_compress_mem_to_mem(void *pOut_buf, size_t out_buf_len, + const void *pSrc_buf, size_t src_buf_len, + int flags) { + tdefl_output_buffer out_buf; + MZ_CLEAR_OBJ(out_buf); + if (!pOut_buf) return 0; + out_buf.m_pBuf = (mz_uint8 *)pOut_buf; + out_buf.m_capacity = out_buf_len; + if (!tdefl_compress_mem_to_output( + pSrc_buf, src_buf_len, tdefl_output_buffer_putter, &out_buf, flags)) + return 0; + return out_buf.m_size; +} + +#ifndef MINIZ_NO_ZLIB_APIS +static const mz_uint s_tdefl_num_probes[11] = {0, 1, 6, 32, 16, 32, + 128, 256, 512, 768, 1500}; + +// level may actually range from [0,10] (10 is a "hidden" max level, where we +// want a bit more compression and it's fine if throughput to fall off a cliff +// on some files). +mz_uint tdefl_create_comp_flags_from_zip_params(int level, int window_bits, + int strategy) { + mz_uint comp_flags = + s_tdefl_num_probes[(level >= 0) ? MZ_MIN(10, level) : MZ_DEFAULT_LEVEL] | + ((level <= 3) ? TDEFL_GREEDY_PARSING_FLAG : 0); + if (window_bits > 0) comp_flags |= TDEFL_WRITE_ZLIB_HEADER; + + if (!level) + comp_flags |= TDEFL_FORCE_ALL_RAW_BLOCKS; + else if (strategy == MZ_FILTERED) + comp_flags |= TDEFL_FILTER_MATCHES; + else if (strategy == MZ_HUFFMAN_ONLY) + comp_flags &= ~TDEFL_MAX_PROBES_MASK; + else if (strategy == MZ_FIXED) + comp_flags |= TDEFL_FORCE_ALL_STATIC_BLOCKS; + else if (strategy == MZ_RLE) + comp_flags |= TDEFL_RLE_MATCHES; + + return comp_flags; +} +#endif // MINIZ_NO_ZLIB_APIS + +#ifdef _MSC_VER +#pragma warning(push) +#pragma warning(disable : 4204) // nonstandard extension used : non-constant + // aggregate initializer (also supported by GNU + // C and C99, so no big deal) +#pragma warning(disable : 4244) // 'initializing': conversion from '__int64' to + // 'int', possible loss of data +#pragma warning(disable : 4267) // 'argument': conversion from '__int64' to + // 'int', possible loss of data +#pragma warning(disable : 4996) // 'strdup': The POSIX name for this item is + // deprecated. Instead, use the ISO C and C++ + // conformant name: _strdup. +#endif + +// Simple PNG writer function by Alex Evans, 2011. Released into the public +// domain: https://gist.github.com/908299, more context at +// http://altdevblogaday.org/2011/04/06/a-smaller-jpg-encoder/. +// This is actually a modification of Alex's original code so PNG files +// generated by this function pass pngcheck. +void *tdefl_write_image_to_png_file_in_memory_ex(const void *pImage, int w, + int h, int num_chans, + size_t *pLen_out, + mz_uint level, mz_bool flip) { + // Using a local copy of this array here in case MINIZ_NO_ZLIB_APIS was + // defined. + static const mz_uint s_tdefl_png_num_probes[11] = { + 0, 1, 6, 32, 16, 32, 128, 256, 512, 768, 1500}; + tdefl_compressor *pComp = + (tdefl_compressor *)MZ_MALLOC(sizeof(tdefl_compressor)); + tdefl_output_buffer out_buf; + int i, bpl = w * num_chans, y, z; + mz_uint32 c; + *pLen_out = 0; + if (!pComp) return NULL; + MZ_CLEAR_OBJ(out_buf); + out_buf.m_expandable = MZ_TRUE; + out_buf.m_capacity = 57 + MZ_MAX(64, (1 + bpl) * h); + if (NULL == (out_buf.m_pBuf = (mz_uint8 *)MZ_MALLOC(out_buf.m_capacity))) { + MZ_FREE(pComp); + return NULL; + } + // write dummy header + for (z = 41; z; --z) tdefl_output_buffer_putter(&z, 1, &out_buf); + // compress image data + tdefl_init( + pComp, tdefl_output_buffer_putter, &out_buf, + s_tdefl_png_num_probes[MZ_MIN(10, level)] | TDEFL_WRITE_ZLIB_HEADER); + for (y = 0; y < h; ++y) { + tdefl_compress_buffer(pComp, &z, 1, TDEFL_NO_FLUSH); + tdefl_compress_buffer(pComp, + (mz_uint8 *)pImage + (flip ? (h - 1 - y) : y) * bpl, + bpl, TDEFL_NO_FLUSH); + } + if (tdefl_compress_buffer(pComp, NULL, 0, TDEFL_FINISH) != + TDEFL_STATUS_DONE) { + MZ_FREE(pComp); + MZ_FREE(out_buf.m_pBuf); + return NULL; + } + // write real header + *pLen_out = out_buf.m_size - 41; + { + static const mz_uint8 chans[] = {0x00, 0x00, 0x04, 0x02, 0x06}; + mz_uint8 pnghdr[41] = {0x89, + 0x50, + 0x4e, + 0x47, + 0x0d, + 0x0a, + 0x1a, + 0x0a, + 0x00, + 0x00, + 0x00, + 0x0d, + 0x49, + 0x48, + 0x44, + 0x52, + 0, + 0, + (mz_uint8)(w >> 8), + (mz_uint8)w, + 0, + 0, + (mz_uint8)(h >> 8), + (mz_uint8)h, + 8, + chans[num_chans], + 0, + 0, + 0, + 0, + 0, + 0, + 0, + (mz_uint8)(*pLen_out >> 24), + (mz_uint8)(*pLen_out >> 16), + (mz_uint8)(*pLen_out >> 8), + (mz_uint8)*pLen_out, + 0x49, + 0x44, + 0x41, + 0x54}; + c = (mz_uint32)mz_crc32(MZ_CRC32_INIT, pnghdr + 12, 17); + for (i = 0; i < 4; ++i, c <<= 8) + ((mz_uint8 *)(pnghdr + 29))[i] = (mz_uint8)(c >> 24); + memcpy(out_buf.m_pBuf, pnghdr, 41); + } + // write footer (IDAT CRC-32, followed by IEND chunk) + if (!tdefl_output_buffer_putter( + "\0\0\0\0\0\0\0\0\x49\x45\x4e\x44\xae\x42\x60\x82", 16, &out_buf)) { + *pLen_out = 0; + MZ_FREE(pComp); + MZ_FREE(out_buf.m_pBuf); + return NULL; + } + c = (mz_uint32)mz_crc32(MZ_CRC32_INIT, out_buf.m_pBuf + 41 - 4, + *pLen_out + 4); + for (i = 0; i < 4; ++i, c <<= 8) + (out_buf.m_pBuf + out_buf.m_size - 16)[i] = (mz_uint8)(c >> 24); + // compute final size of file, grab compressed data buffer and return + *pLen_out += 57; + MZ_FREE(pComp); + return out_buf.m_pBuf; +} +void *tdefl_write_image_to_png_file_in_memory(const void *pImage, int w, int h, + int num_chans, size_t *pLen_out) { + // Level 6 corresponds to TDEFL_DEFAULT_MAX_PROBES or MZ_DEFAULT_LEVEL (but we + // can't depend on MZ_DEFAULT_LEVEL being available in case the zlib API's + // where #defined out) + return tdefl_write_image_to_png_file_in_memory_ex(pImage, w, h, num_chans, + pLen_out, 6, MZ_FALSE); +} + +// ------------------- .ZIP archive reading + +#ifndef MINIZ_NO_ARCHIVE_APIS +#error "No arvhive APIs" + +#ifdef MINIZ_NO_STDIO +#define MZ_FILE void * +#else +#include +#include + +#if defined(_MSC_VER) || defined(__MINGW64__) +static FILE *mz_fopen(const char *pFilename, const char *pMode) { + FILE *pFile = NULL; + fopen_s(&pFile, pFilename, pMode); + return pFile; +} +static FILE *mz_freopen(const char *pPath, const char *pMode, FILE *pStream) { + FILE *pFile = NULL; + if (freopen_s(&pFile, pPath, pMode, pStream)) return NULL; + return pFile; +} +#ifndef MINIZ_NO_TIME +#include +#endif +#define MZ_FILE FILE +#define MZ_FOPEN mz_fopen +#define MZ_FCLOSE fclose +#define MZ_FREAD fread +#define MZ_FWRITE fwrite +#define MZ_FTELL64 _ftelli64 +#define MZ_FSEEK64 _fseeki64 +#define MZ_FILE_STAT_STRUCT _stat +#define MZ_FILE_STAT _stat +#define MZ_FFLUSH fflush +#define MZ_FREOPEN mz_freopen +#define MZ_DELETE_FILE remove +#elif defined(__MINGW32__) +#ifndef MINIZ_NO_TIME +#include +#endif +#define MZ_FILE FILE +#define MZ_FOPEN(f, m) fopen(f, m) +#define MZ_FCLOSE fclose +#define MZ_FREAD fread +#define MZ_FWRITE fwrite +#define MZ_FTELL64 ftello64 +#define MZ_FSEEK64 fseeko64 +#define MZ_FILE_STAT_STRUCT _stat +#define MZ_FILE_STAT _stat +#define MZ_FFLUSH fflush +#define MZ_FREOPEN(f, m, s) freopen(f, m, s) +#define MZ_DELETE_FILE remove +#elif defined(__TINYC__) +#ifndef MINIZ_NO_TIME +#include +#endif +#define MZ_FILE FILE +#define MZ_FOPEN(f, m) fopen(f, m) +#define MZ_FCLOSE fclose +#define MZ_FREAD fread +#define MZ_FWRITE fwrite +#define MZ_FTELL64 ftell +#define MZ_FSEEK64 fseek +#define MZ_FILE_STAT_STRUCT stat +#define MZ_FILE_STAT stat +#define MZ_FFLUSH fflush +#define MZ_FREOPEN(f, m, s) freopen(f, m, s) +#define MZ_DELETE_FILE remove +#elif defined(__GNUC__) && defined(_LARGEFILE64_SOURCE) && _LARGEFILE64_SOURCE +#ifndef MINIZ_NO_TIME +#include +#endif +#define MZ_FILE FILE +#define MZ_FOPEN(f, m) fopen64(f, m) +#define MZ_FCLOSE fclose +#define MZ_FREAD fread +#define MZ_FWRITE fwrite +#define MZ_FTELL64 ftello64 +#define MZ_FSEEK64 fseeko64 +#define MZ_FILE_STAT_STRUCT stat64 +#define MZ_FILE_STAT stat64 +#define MZ_FFLUSH fflush +#define MZ_FREOPEN(p, m, s) freopen64(p, m, s) +#define MZ_DELETE_FILE remove +#else +#ifndef MINIZ_NO_TIME +#include +#endif +#define MZ_FILE FILE +#define MZ_FOPEN(f, m) fopen(f, m) +#define MZ_FCLOSE fclose +#define MZ_FREAD fread +#define MZ_FWRITE fwrite +#define MZ_FTELL64 ftello +#define MZ_FSEEK64 fseeko +#define MZ_FILE_STAT_STRUCT stat +#define MZ_FILE_STAT stat +#define MZ_FFLUSH fflush +#define MZ_FREOPEN(f, m, s) freopen(f, m, s) +#define MZ_DELETE_FILE remove +#endif // #ifdef _MSC_VER +#endif // #ifdef MINIZ_NO_STDIO + +#define MZ_TOLOWER(c) ((((c) >= 'A') && ((c) <= 'Z')) ? ((c) - 'A' + 'a') : (c)) + +// Various ZIP archive enums. To completely avoid cross platform compiler +// alignment and platform endian issues, miniz.c doesn't use structs for any of +// this stuff. +enum { + // ZIP archive identifiers and record sizes + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIG = 0x06054b50, + MZ_ZIP_CENTRAL_DIR_HEADER_SIG = 0x02014b50, + MZ_ZIP_LOCAL_DIR_HEADER_SIG = 0x04034b50, + MZ_ZIP_LOCAL_DIR_HEADER_SIZE = 30, + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE = 46, + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE = 22, + // Central directory header record offsets + MZ_ZIP_CDH_SIG_OFS = 0, + MZ_ZIP_CDH_VERSION_MADE_BY_OFS = 4, + MZ_ZIP_CDH_VERSION_NEEDED_OFS = 6, + MZ_ZIP_CDH_BIT_FLAG_OFS = 8, + MZ_ZIP_CDH_METHOD_OFS = 10, + MZ_ZIP_CDH_FILE_TIME_OFS = 12, + MZ_ZIP_CDH_FILE_DATE_OFS = 14, + MZ_ZIP_CDH_CRC32_OFS = 16, + MZ_ZIP_CDH_COMPRESSED_SIZE_OFS = 20, + MZ_ZIP_CDH_DECOMPRESSED_SIZE_OFS = 24, + MZ_ZIP_CDH_FILENAME_LEN_OFS = 28, + MZ_ZIP_CDH_EXTRA_LEN_OFS = 30, + MZ_ZIP_CDH_COMMENT_LEN_OFS = 32, + MZ_ZIP_CDH_DISK_START_OFS = 34, + MZ_ZIP_CDH_INTERNAL_ATTR_OFS = 36, + MZ_ZIP_CDH_EXTERNAL_ATTR_OFS = 38, + MZ_ZIP_CDH_LOCAL_HEADER_OFS = 42, + // Local directory header offsets + MZ_ZIP_LDH_SIG_OFS = 0, + MZ_ZIP_LDH_VERSION_NEEDED_OFS = 4, + MZ_ZIP_LDH_BIT_FLAG_OFS = 6, + MZ_ZIP_LDH_METHOD_OFS = 8, + MZ_ZIP_LDH_FILE_TIME_OFS = 10, + MZ_ZIP_LDH_FILE_DATE_OFS = 12, + MZ_ZIP_LDH_CRC32_OFS = 14, + MZ_ZIP_LDH_COMPRESSED_SIZE_OFS = 18, + MZ_ZIP_LDH_DECOMPRESSED_SIZE_OFS = 22, + MZ_ZIP_LDH_FILENAME_LEN_OFS = 26, + MZ_ZIP_LDH_EXTRA_LEN_OFS = 28, + // End of central directory offsets + MZ_ZIP_ECDH_SIG_OFS = 0, + MZ_ZIP_ECDH_NUM_THIS_DISK_OFS = 4, + MZ_ZIP_ECDH_NUM_DISK_CDIR_OFS = 6, + MZ_ZIP_ECDH_CDIR_NUM_ENTRIES_ON_DISK_OFS = 8, + MZ_ZIP_ECDH_CDIR_TOTAL_ENTRIES_OFS = 10, + MZ_ZIP_ECDH_CDIR_SIZE_OFS = 12, + MZ_ZIP_ECDH_CDIR_OFS_OFS = 16, + MZ_ZIP_ECDH_COMMENT_SIZE_OFS = 20, +}; + +typedef struct { + void *m_p; + size_t m_size, m_capacity; + mz_uint m_element_size; +} mz_zip_array; + +struct mz_zip_internal_state_tag { + mz_zip_array m_central_dir; + mz_zip_array m_central_dir_offsets; + mz_zip_array m_sorted_central_dir_offsets; + MZ_FILE *m_pFile; + void *m_pMem; + size_t m_mem_size; + size_t m_mem_capacity; +}; + +#define MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(array_ptr, element_size) \ + (array_ptr)->m_element_size = element_size +#define MZ_ZIP_ARRAY_ELEMENT(array_ptr, element_type, index) \ + ((element_type *)((array_ptr)->m_p))[index] + +static MZ_FORCEINLINE void mz_zip_array_clear(mz_zip_archive *pZip, + mz_zip_array *pArray) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pArray->m_p); + memset(pArray, 0, sizeof(mz_zip_array)); +} + +static mz_bool mz_zip_array_ensure_capacity(mz_zip_archive *pZip, + mz_zip_array *pArray, + size_t min_new_capacity, + mz_uint growing) { + void *pNew_p; + size_t new_capacity = min_new_capacity; + MZ_ASSERT(pArray->m_element_size); + if (pArray->m_capacity >= min_new_capacity) return MZ_TRUE; + if (growing) { + new_capacity = MZ_MAX(1, pArray->m_capacity); + while (new_capacity < min_new_capacity) new_capacity *= 2; + } + if (NULL == (pNew_p = pZip->m_pRealloc(pZip->m_pAlloc_opaque, pArray->m_p, + pArray->m_element_size, new_capacity))) + return MZ_FALSE; + pArray->m_p = pNew_p; + pArray->m_capacity = new_capacity; + return MZ_TRUE; +} + +static MZ_FORCEINLINE mz_bool mz_zip_array_reserve(mz_zip_archive *pZip, + mz_zip_array *pArray, + size_t new_capacity, + mz_uint growing) { + if (new_capacity > pArray->m_capacity) { + if (!mz_zip_array_ensure_capacity(pZip, pArray, new_capacity, growing)) + return MZ_FALSE; + } + return MZ_TRUE; +} + +static MZ_FORCEINLINE mz_bool mz_zip_array_resize(mz_zip_archive *pZip, + mz_zip_array *pArray, + size_t new_size, + mz_uint growing) { + if (new_size > pArray->m_capacity) { + if (!mz_zip_array_ensure_capacity(pZip, pArray, new_size, growing)) + return MZ_FALSE; + } + pArray->m_size = new_size; + return MZ_TRUE; +} + +static MZ_FORCEINLINE mz_bool mz_zip_array_ensure_room(mz_zip_archive *pZip, + mz_zip_array *pArray, + size_t n) { + return mz_zip_array_reserve(pZip, pArray, pArray->m_size + n, MZ_TRUE); +} + +static MZ_FORCEINLINE mz_bool mz_zip_array_push_back(mz_zip_archive *pZip, + mz_zip_array *pArray, + const void *pElements, + size_t n) { + size_t orig_size = pArray->m_size; + if (!mz_zip_array_resize(pZip, pArray, orig_size + n, MZ_TRUE)) + return MZ_FALSE; + memcpy((mz_uint8 *)pArray->m_p + orig_size * pArray->m_element_size, + pElements, n * pArray->m_element_size); + return MZ_TRUE; +} + +#ifndef MINIZ_NO_TIME +static time_t mz_zip_dos_to_time_t(int dos_time, int dos_date) { + struct tm tm; + memset(&tm, 0, sizeof(tm)); + tm.tm_isdst = -1; + tm.tm_year = ((dos_date >> 9) & 127) + 1980 - 1900; + tm.tm_mon = ((dos_date >> 5) & 15) - 1; + tm.tm_mday = dos_date & 31; + tm.tm_hour = (dos_time >> 11) & 31; + tm.tm_min = (dos_time >> 5) & 63; + tm.tm_sec = (dos_time << 1) & 62; + return mktime(&tm); +} + +static void mz_zip_time_to_dos_time(time_t time, mz_uint16 *pDOS_time, + mz_uint16 *pDOS_date) { +#ifdef _MSC_VER + struct tm tm_struct; + struct tm *tm = &tm_struct; + errno_t err = localtime_s(tm, &time); + if (err) { + *pDOS_date = 0; + *pDOS_time = 0; + return; + } +#else + struct tm *tm = localtime(&time); +#endif + *pDOS_time = (mz_uint16)(((tm->tm_hour) << 11) + ((tm->tm_min) << 5) + + ((tm->tm_sec) >> 1)); + *pDOS_date = (mz_uint16)(((tm->tm_year + 1900 - 1980) << 9) + + ((tm->tm_mon + 1) << 5) + tm->tm_mday); +} +#endif + +#ifndef MINIZ_NO_STDIO +static mz_bool mz_zip_get_file_modified_time(const char *pFilename, + mz_uint16 *pDOS_time, + mz_uint16 *pDOS_date) { +#ifdef MINIZ_NO_TIME + (void)pFilename; + *pDOS_date = *pDOS_time = 0; +#else + struct MZ_FILE_STAT_STRUCT file_stat; + // On Linux with x86 glibc, this call will fail on large files (>= 0x80000000 + // bytes) unless you compiled with _LARGEFILE64_SOURCE. Argh. + if (MZ_FILE_STAT(pFilename, &file_stat) != 0) return MZ_FALSE; + mz_zip_time_to_dos_time(file_stat.st_mtime, pDOS_time, pDOS_date); +#endif // #ifdef MINIZ_NO_TIME + return MZ_TRUE; +} + +#ifndef MINIZ_NO_TIME +static mz_bool mz_zip_set_file_times(const char *pFilename, time_t access_time, + time_t modified_time) { + struct utimbuf t; + t.actime = access_time; + t.modtime = modified_time; + return !utime(pFilename, &t); +} +#endif // #ifndef MINIZ_NO_TIME +#endif // #ifndef MINIZ_NO_STDIO + +static mz_bool mz_zip_reader_init_internal(mz_zip_archive *pZip, + mz_uint32 flags) { + (void)flags; + if ((!pZip) || (pZip->m_pState) || (pZip->m_zip_mode != MZ_ZIP_MODE_INVALID)) + return MZ_FALSE; + + if (!pZip->m_pAlloc) pZip->m_pAlloc = def_alloc_func; + if (!pZip->m_pFree) pZip->m_pFree = def_free_func; + if (!pZip->m_pRealloc) pZip->m_pRealloc = def_realloc_func; + + pZip->m_zip_mode = MZ_ZIP_MODE_READING; + pZip->m_archive_size = 0; + pZip->m_central_directory_file_ofs = 0; + pZip->m_total_files = 0; + + if (NULL == (pZip->m_pState = (mz_zip_internal_state *)pZip->m_pAlloc( + pZip->m_pAlloc_opaque, 1, sizeof(mz_zip_internal_state)))) + return MZ_FALSE; + memset(pZip->m_pState, 0, sizeof(mz_zip_internal_state)); + MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(&pZip->m_pState->m_central_dir, + sizeof(mz_uint8)); + MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(&pZip->m_pState->m_central_dir_offsets, + sizeof(mz_uint32)); + MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(&pZip->m_pState->m_sorted_central_dir_offsets, + sizeof(mz_uint32)); + return MZ_TRUE; +} + +static MZ_FORCEINLINE mz_bool +mz_zip_reader_filename_less(const mz_zip_array *pCentral_dir_array, + const mz_zip_array *pCentral_dir_offsets, + mz_uint l_index, mz_uint r_index) { + const mz_uint8 *pL = &MZ_ZIP_ARRAY_ELEMENT( + pCentral_dir_array, mz_uint8, + MZ_ZIP_ARRAY_ELEMENT(pCentral_dir_offsets, mz_uint32, + l_index)), + *pE; + const mz_uint8 *pR = &MZ_ZIP_ARRAY_ELEMENT( + pCentral_dir_array, mz_uint8, + MZ_ZIP_ARRAY_ELEMENT(pCentral_dir_offsets, mz_uint32, r_index)); + mz_uint l_len = MZ_READ_LE16(pL + MZ_ZIP_CDH_FILENAME_LEN_OFS), + r_len = MZ_READ_LE16(pR + MZ_ZIP_CDH_FILENAME_LEN_OFS); + mz_uint8 l = 0, r = 0; + pL += MZ_ZIP_CENTRAL_DIR_HEADER_SIZE; + pR += MZ_ZIP_CENTRAL_DIR_HEADER_SIZE; + pE = pL + MZ_MIN(l_len, r_len); + while (pL < pE) { + if ((l = MZ_TOLOWER(*pL)) != (r = MZ_TOLOWER(*pR))) break; + pL++; + pR++; + } + return (pL == pE) ? (l_len < r_len) : (l < r); +} + +#define MZ_SWAP_UINT32(a, b) \ + do { \ + mz_uint32 t = a; \ + a = b; \ + b = t; \ + } \ + MZ_MACRO_END + +// Heap sort of lowercased filenames, used to help accelerate plain central +// directory searches by mz_zip_reader_locate_file(). (Could also use qsort(), +// but it could allocate memory.) +static void mz_zip_reader_sort_central_dir_offsets_by_filename( + mz_zip_archive *pZip) { + mz_zip_internal_state *pState = pZip->m_pState; + const mz_zip_array *pCentral_dir_offsets = &pState->m_central_dir_offsets; + const mz_zip_array *pCentral_dir = &pState->m_central_dir; + mz_uint32 *pIndices = &MZ_ZIP_ARRAY_ELEMENT( + &pState->m_sorted_central_dir_offsets, mz_uint32, 0); + const int size = pZip->m_total_files; + int start = (size - 2) >> 1, end; + while (start >= 0) { + int child, root = start; + for (;;) { + if ((child = (root << 1) + 1) >= size) break; + child += + (((child + 1) < size) && + (mz_zip_reader_filename_less(pCentral_dir, pCentral_dir_offsets, + pIndices[child], pIndices[child + 1]))); + if (!mz_zip_reader_filename_less(pCentral_dir, pCentral_dir_offsets, + pIndices[root], pIndices[child])) + break; + MZ_SWAP_UINT32(pIndices[root], pIndices[child]); + root = child; + } + start--; + } + + end = size - 1; + while (end > 0) { + int child, root = 0; + MZ_SWAP_UINT32(pIndices[end], pIndices[0]); + for (;;) { + if ((child = (root << 1) + 1) >= end) break; + child += + (((child + 1) < end) && + mz_zip_reader_filename_less(pCentral_dir, pCentral_dir_offsets, + pIndices[child], pIndices[child + 1])); + if (!mz_zip_reader_filename_less(pCentral_dir, pCentral_dir_offsets, + pIndices[root], pIndices[child])) + break; + MZ_SWAP_UINT32(pIndices[root], pIndices[child]); + root = child; + } + end--; + } +} + +static mz_bool mz_zip_reader_read_central_dir(mz_zip_archive *pZip, + mz_uint32 flags) { + mz_uint cdir_size, num_this_disk, cdir_disk_index; + mz_uint64 cdir_ofs; + mz_int64 cur_file_ofs; + const mz_uint8 *p; + mz_uint32 buf_u32[4096 / sizeof(mz_uint32)]; + mz_uint8 *pBuf = (mz_uint8 *)buf_u32; + mz_bool sort_central_dir = + ((flags & MZ_ZIP_FLAG_DO_NOT_SORT_CENTRAL_DIRECTORY) == 0); + // Basic sanity checks - reject files which are too small, and check the first + // 4 bytes of the file to make sure a local header is there. + if (pZip->m_archive_size < MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE) + return MZ_FALSE; + // Find the end of central directory record by scanning the file from the end + // towards the beginning. + cur_file_ofs = + MZ_MAX((mz_int64)pZip->m_archive_size - (mz_int64)sizeof(buf_u32), 0); + for (;;) { + int i, + n = (int)MZ_MIN(sizeof(buf_u32), pZip->m_archive_size - cur_file_ofs); + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pBuf, n) != (mz_uint)n) + return MZ_FALSE; + for (i = n - 4; i >= 0; --i) + if (MZ_READ_LE32(pBuf + i) == MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIG) break; + if (i >= 0) { + cur_file_ofs += i; + break; + } + if ((!cur_file_ofs) || ((pZip->m_archive_size - cur_file_ofs) >= + (0xFFFF + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE))) + return MZ_FALSE; + cur_file_ofs = MZ_MAX(cur_file_ofs - (sizeof(buf_u32) - 3), 0); + } + // Read and verify the end of central directory record. + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pBuf, + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE) != + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE) + return MZ_FALSE; + if ((MZ_READ_LE32(pBuf + MZ_ZIP_ECDH_SIG_OFS) != + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIG) || + ((pZip->m_total_files = + MZ_READ_LE16(pBuf + MZ_ZIP_ECDH_CDIR_TOTAL_ENTRIES_OFS)) != + MZ_READ_LE16(pBuf + MZ_ZIP_ECDH_CDIR_NUM_ENTRIES_ON_DISK_OFS))) + return MZ_FALSE; + + num_this_disk = MZ_READ_LE16(pBuf + MZ_ZIP_ECDH_NUM_THIS_DISK_OFS); + cdir_disk_index = MZ_READ_LE16(pBuf + MZ_ZIP_ECDH_NUM_DISK_CDIR_OFS); + if (((num_this_disk | cdir_disk_index) != 0) && + ((num_this_disk != 1) || (cdir_disk_index != 1))) + return MZ_FALSE; + + if ((cdir_size = MZ_READ_LE32(pBuf + MZ_ZIP_ECDH_CDIR_SIZE_OFS)) < + pZip->m_total_files * MZ_ZIP_CENTRAL_DIR_HEADER_SIZE) + return MZ_FALSE; + + cdir_ofs = MZ_READ_LE32(pBuf + MZ_ZIP_ECDH_CDIR_OFS_OFS); + if ((cdir_ofs + (mz_uint64)cdir_size) > pZip->m_archive_size) return MZ_FALSE; + + pZip->m_central_directory_file_ofs = cdir_ofs; + + if (pZip->m_total_files) { + mz_uint i, n; + + // Read the entire central directory into a heap block, and allocate another + // heap block to hold the unsorted central dir file record offsets, and + // another to hold the sorted indices. + if ((!mz_zip_array_resize(pZip, &pZip->m_pState->m_central_dir, cdir_size, + MZ_FALSE)) || + (!mz_zip_array_resize(pZip, &pZip->m_pState->m_central_dir_offsets, + pZip->m_total_files, MZ_FALSE))) + return MZ_FALSE; + + if (sort_central_dir) { + if (!mz_zip_array_resize(pZip, + &pZip->m_pState->m_sorted_central_dir_offsets, + pZip->m_total_files, MZ_FALSE)) + return MZ_FALSE; + } + + if (pZip->m_pRead(pZip->m_pIO_opaque, cdir_ofs, + pZip->m_pState->m_central_dir.m_p, + cdir_size) != cdir_size) + return MZ_FALSE; + + // Now create an index into the central directory file records, do some + // basic sanity checking on each record, and check for zip64 entries (which + // are not yet supported). + p = (const mz_uint8 *)pZip->m_pState->m_central_dir.m_p; + for (n = cdir_size, i = 0; i < pZip->m_total_files; ++i) { + mz_uint total_header_size, comp_size, decomp_size, disk_index; + if ((n < MZ_ZIP_CENTRAL_DIR_HEADER_SIZE) || + (MZ_READ_LE32(p) != MZ_ZIP_CENTRAL_DIR_HEADER_SIG)) + return MZ_FALSE; + MZ_ZIP_ARRAY_ELEMENT(&pZip->m_pState->m_central_dir_offsets, mz_uint32, + i) = + (mz_uint32)(p - (const mz_uint8 *)pZip->m_pState->m_central_dir.m_p); + if (sort_central_dir) + MZ_ZIP_ARRAY_ELEMENT(&pZip->m_pState->m_sorted_central_dir_offsets, + mz_uint32, i) = i; + comp_size = MZ_READ_LE32(p + MZ_ZIP_CDH_COMPRESSED_SIZE_OFS); + decomp_size = MZ_READ_LE32(p + MZ_ZIP_CDH_DECOMPRESSED_SIZE_OFS); + if (((!MZ_READ_LE32(p + MZ_ZIP_CDH_METHOD_OFS)) && + (decomp_size != comp_size)) || + (decomp_size && !comp_size) || (decomp_size == 0xFFFFFFFF) || + (comp_size == 0xFFFFFFFF)) + return MZ_FALSE; + disk_index = MZ_READ_LE16(p + MZ_ZIP_CDH_DISK_START_OFS); + if ((disk_index != num_this_disk) && (disk_index != 1)) return MZ_FALSE; + if (((mz_uint64)MZ_READ_LE32(p + MZ_ZIP_CDH_LOCAL_HEADER_OFS) + + MZ_ZIP_LOCAL_DIR_HEADER_SIZE + comp_size) > pZip->m_archive_size) + return MZ_FALSE; + if ((total_header_size = MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + + MZ_READ_LE16(p + MZ_ZIP_CDH_FILENAME_LEN_OFS) + + MZ_READ_LE16(p + MZ_ZIP_CDH_EXTRA_LEN_OFS) + + MZ_READ_LE16(p + MZ_ZIP_CDH_COMMENT_LEN_OFS)) > + n) + return MZ_FALSE; + n -= total_header_size; + p += total_header_size; + } + } + + if (sort_central_dir) + mz_zip_reader_sort_central_dir_offsets_by_filename(pZip); + + return MZ_TRUE; +} + +mz_bool mz_zip_reader_init(mz_zip_archive *pZip, mz_uint64 size, + mz_uint32 flags) { + if ((!pZip) || (!pZip->m_pRead)) return MZ_FALSE; + if (!mz_zip_reader_init_internal(pZip, flags)) return MZ_FALSE; + pZip->m_archive_size = size; + if (!mz_zip_reader_read_central_dir(pZip, flags)) { + mz_zip_reader_end(pZip); + return MZ_FALSE; + } + return MZ_TRUE; +} + +static size_t mz_zip_mem_read_func(void *pOpaque, mz_uint64 file_ofs, + void *pBuf, size_t n) { + mz_zip_archive *pZip = (mz_zip_archive *)pOpaque; + size_t s = (file_ofs >= pZip->m_archive_size) + ? 0 + : (size_t)MZ_MIN(pZip->m_archive_size - file_ofs, n); + memcpy(pBuf, (const mz_uint8 *)pZip->m_pState->m_pMem + file_ofs, s); + return s; +} + +mz_bool mz_zip_reader_init_mem(mz_zip_archive *pZip, const void *pMem, + size_t size, mz_uint32 flags) { + if (!mz_zip_reader_init_internal(pZip, flags)) return MZ_FALSE; + pZip->m_archive_size = size; + pZip->m_pRead = mz_zip_mem_read_func; + pZip->m_pIO_opaque = pZip; +#ifdef __cplusplus + pZip->m_pState->m_pMem = const_cast(pMem); +#else + pZip->m_pState->m_pMem = (void *)pMem; +#endif + pZip->m_pState->m_mem_size = size; + if (!mz_zip_reader_read_central_dir(pZip, flags)) { + mz_zip_reader_end(pZip); + return MZ_FALSE; + } + return MZ_TRUE; +} + +#ifndef MINIZ_NO_STDIO +static size_t mz_zip_file_read_func(void *pOpaque, mz_uint64 file_ofs, + void *pBuf, size_t n) { + mz_zip_archive *pZip = (mz_zip_archive *)pOpaque; + mz_int64 cur_ofs = MZ_FTELL64(pZip->m_pState->m_pFile); + if (((mz_int64)file_ofs < 0) || + (((cur_ofs != (mz_int64)file_ofs)) && + (MZ_FSEEK64(pZip->m_pState->m_pFile, (mz_int64)file_ofs, SEEK_SET)))) + return 0; + return MZ_FREAD(pBuf, 1, n, pZip->m_pState->m_pFile); +} + +mz_bool mz_zip_reader_init_file(mz_zip_archive *pZip, const char *pFilename, + mz_uint32 flags) { + mz_uint64 file_size; + MZ_FILE *pFile = MZ_FOPEN(pFilename, "rb"); + if (!pFile) return MZ_FALSE; + if (MZ_FSEEK64(pFile, 0, SEEK_END)) { + MZ_FCLOSE(pFile); + return MZ_FALSE; + } + file_size = MZ_FTELL64(pFile); + if (!mz_zip_reader_init_internal(pZip, flags)) { + MZ_FCLOSE(pFile); + return MZ_FALSE; + } + pZip->m_pRead = mz_zip_file_read_func; + pZip->m_pIO_opaque = pZip; + pZip->m_pState->m_pFile = pFile; + pZip->m_archive_size = file_size; + if (!mz_zip_reader_read_central_dir(pZip, flags)) { + mz_zip_reader_end(pZip); + return MZ_FALSE; + } + return MZ_TRUE; +} +#endif // #ifndef MINIZ_NO_STDIO + +mz_uint mz_zip_reader_get_num_files(mz_zip_archive *pZip) { + return pZip ? pZip->m_total_files : 0; +} + +static MZ_FORCEINLINE const mz_uint8 *mz_zip_reader_get_cdh( + mz_zip_archive *pZip, mz_uint file_index) { + if ((!pZip) || (!pZip->m_pState) || (file_index >= pZip->m_total_files) || + (pZip->m_zip_mode != MZ_ZIP_MODE_READING)) + return NULL; + return &MZ_ZIP_ARRAY_ELEMENT( + &pZip->m_pState->m_central_dir, mz_uint8, + MZ_ZIP_ARRAY_ELEMENT(&pZip->m_pState->m_central_dir_offsets, mz_uint32, + file_index)); +} + +mz_bool mz_zip_reader_is_file_encrypted(mz_zip_archive *pZip, + mz_uint file_index) { + mz_uint m_bit_flag; + const mz_uint8 *p = mz_zip_reader_get_cdh(pZip, file_index); + if (!p) return MZ_FALSE; + m_bit_flag = MZ_READ_LE16(p + MZ_ZIP_CDH_BIT_FLAG_OFS); + return (m_bit_flag & 1); +} + +mz_bool mz_zip_reader_is_file_a_directory(mz_zip_archive *pZip, + mz_uint file_index) { + mz_uint filename_len, external_attr; + const mz_uint8 *p = mz_zip_reader_get_cdh(pZip, file_index); + if (!p) return MZ_FALSE; + + // First see if the filename ends with a '/' character. + filename_len = MZ_READ_LE16(p + MZ_ZIP_CDH_FILENAME_LEN_OFS); + if (filename_len) { + if (*(p + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + filename_len - 1) == '/') + return MZ_TRUE; + } + + // Bugfix: This code was also checking if the internal attribute was non-zero, + // which wasn't correct. + // Most/all zip writers (hopefully) set DOS file/directory attributes in the + // low 16-bits, so check for the DOS directory flag and ignore the source OS + // ID in the created by field. + // FIXME: Remove this check? Is it necessary - we already check the filename. + external_attr = MZ_READ_LE32(p + MZ_ZIP_CDH_EXTERNAL_ATTR_OFS); + if ((external_attr & 0x10) != 0) return MZ_TRUE; + + return MZ_FALSE; +} + +mz_bool mz_zip_reader_file_stat(mz_zip_archive *pZip, mz_uint file_index, + mz_zip_archive_file_stat *pStat) { + mz_uint n; + const mz_uint8 *p = mz_zip_reader_get_cdh(pZip, file_index); + if ((!p) || (!pStat)) return MZ_FALSE; + + // Unpack the central directory record. + pStat->m_file_index = file_index; + pStat->m_central_dir_ofs = MZ_ZIP_ARRAY_ELEMENT( + &pZip->m_pState->m_central_dir_offsets, mz_uint32, file_index); + pStat->m_version_made_by = MZ_READ_LE16(p + MZ_ZIP_CDH_VERSION_MADE_BY_OFS); + pStat->m_version_needed = MZ_READ_LE16(p + MZ_ZIP_CDH_VERSION_NEEDED_OFS); + pStat->m_bit_flag = MZ_READ_LE16(p + MZ_ZIP_CDH_BIT_FLAG_OFS); + pStat->m_method = MZ_READ_LE16(p + MZ_ZIP_CDH_METHOD_OFS); +#ifndef MINIZ_NO_TIME + pStat->m_time = + mz_zip_dos_to_time_t(MZ_READ_LE16(p + MZ_ZIP_CDH_FILE_TIME_OFS), + MZ_READ_LE16(p + MZ_ZIP_CDH_FILE_DATE_OFS)); +#endif + pStat->m_crc32 = MZ_READ_LE32(p + MZ_ZIP_CDH_CRC32_OFS); + pStat->m_comp_size = MZ_READ_LE32(p + MZ_ZIP_CDH_COMPRESSED_SIZE_OFS); + pStat->m_uncomp_size = MZ_READ_LE32(p + MZ_ZIP_CDH_DECOMPRESSED_SIZE_OFS); + pStat->m_internal_attr = MZ_READ_LE16(p + MZ_ZIP_CDH_INTERNAL_ATTR_OFS); + pStat->m_external_attr = MZ_READ_LE32(p + MZ_ZIP_CDH_EXTERNAL_ATTR_OFS); + pStat->m_local_header_ofs = MZ_READ_LE32(p + MZ_ZIP_CDH_LOCAL_HEADER_OFS); + + // Copy as much of the filename and comment as possible. + n = MZ_READ_LE16(p + MZ_ZIP_CDH_FILENAME_LEN_OFS); + n = MZ_MIN(n, MZ_ZIP_MAX_ARCHIVE_FILENAME_SIZE - 1); + memcpy(pStat->m_filename, p + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE, n); + pStat->m_filename[n] = '\0'; + + n = MZ_READ_LE16(p + MZ_ZIP_CDH_COMMENT_LEN_OFS); + n = MZ_MIN(n, MZ_ZIP_MAX_ARCHIVE_FILE_COMMENT_SIZE - 1); + pStat->m_comment_size = n; + memcpy(pStat->m_comment, + p + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + + MZ_READ_LE16(p + MZ_ZIP_CDH_FILENAME_LEN_OFS) + + MZ_READ_LE16(p + MZ_ZIP_CDH_EXTRA_LEN_OFS), + n); + pStat->m_comment[n] = '\0'; + + return MZ_TRUE; +} + +mz_uint mz_zip_reader_get_filename(mz_zip_archive *pZip, mz_uint file_index, + char *pFilename, mz_uint filename_buf_size) { + mz_uint n; + const mz_uint8 *p = mz_zip_reader_get_cdh(pZip, file_index); + if (!p) { + if (filename_buf_size) pFilename[0] = '\0'; + return 0; + } + n = MZ_READ_LE16(p + MZ_ZIP_CDH_FILENAME_LEN_OFS); + if (filename_buf_size) { + n = MZ_MIN(n, filename_buf_size - 1); + memcpy(pFilename, p + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE, n); + pFilename[n] = '\0'; + } + return n + 1; +} + +static MZ_FORCEINLINE mz_bool mz_zip_reader_string_equal(const char *pA, + const char *pB, + mz_uint len, + mz_uint flags) { + mz_uint i; + if (flags & MZ_ZIP_FLAG_CASE_SENSITIVE) return 0 == memcmp(pA, pB, len); + for (i = 0; i < len; ++i) + if (MZ_TOLOWER(pA[i]) != MZ_TOLOWER(pB[i])) return MZ_FALSE; + return MZ_TRUE; +} + +static MZ_FORCEINLINE int mz_zip_reader_filename_compare( + const mz_zip_array *pCentral_dir_array, + const mz_zip_array *pCentral_dir_offsets, mz_uint l_index, const char *pR, + mz_uint r_len) { + const mz_uint8 *pL = &MZ_ZIP_ARRAY_ELEMENT( + pCentral_dir_array, mz_uint8, + MZ_ZIP_ARRAY_ELEMENT(pCentral_dir_offsets, mz_uint32, + l_index)), + *pE; + mz_uint l_len = MZ_READ_LE16(pL + MZ_ZIP_CDH_FILENAME_LEN_OFS); + mz_uint8 l = 0, r = 0; + pL += MZ_ZIP_CENTRAL_DIR_HEADER_SIZE; + pE = pL + MZ_MIN(l_len, r_len); + while (pL < pE) { + if ((l = MZ_TOLOWER(*pL)) != (r = MZ_TOLOWER(*pR))) break; + pL++; + pR++; + } + return (pL == pE) ? (int)(l_len - r_len) : (l - r); +} + +static int mz_zip_reader_locate_file_binary_search(mz_zip_archive *pZip, + const char *pFilename) { + mz_zip_internal_state *pState = pZip->m_pState; + const mz_zip_array *pCentral_dir_offsets = &pState->m_central_dir_offsets; + const mz_zip_array *pCentral_dir = &pState->m_central_dir; + mz_uint32 *pIndices = &MZ_ZIP_ARRAY_ELEMENT( + &pState->m_sorted_central_dir_offsets, mz_uint32, 0); + const int size = pZip->m_total_files; + const mz_uint filename_len = (mz_uint)strlen(pFilename); + int l = 0, h = size - 1; + while (l <= h) { + int m = (l + h) >> 1, file_index = pIndices[m], + comp = + mz_zip_reader_filename_compare(pCentral_dir, pCentral_dir_offsets, + file_index, pFilename, filename_len); + if (!comp) + return file_index; + else if (comp < 0) + l = m + 1; + else + h = m - 1; + } + return -1; +} + +int mz_zip_reader_locate_file(mz_zip_archive *pZip, const char *pName, + const char *pComment, mz_uint flags) { + mz_uint file_index; + size_t name_len, comment_len; + if ((!pZip) || (!pZip->m_pState) || (!pName) || + (pZip->m_zip_mode != MZ_ZIP_MODE_READING)) + return -1; + if (((flags & (MZ_ZIP_FLAG_IGNORE_PATH | MZ_ZIP_FLAG_CASE_SENSITIVE)) == 0) && + (!pComment) && (pZip->m_pState->m_sorted_central_dir_offsets.m_size)) + return mz_zip_reader_locate_file_binary_search(pZip, pName); + name_len = strlen(pName); + if (name_len > 0xFFFF) return -1; + comment_len = pComment ? strlen(pComment) : 0; + if (comment_len > 0xFFFF) return -1; + for (file_index = 0; file_index < pZip->m_total_files; file_index++) { + const mz_uint8 *pHeader = &MZ_ZIP_ARRAY_ELEMENT( + &pZip->m_pState->m_central_dir, mz_uint8, + MZ_ZIP_ARRAY_ELEMENT(&pZip->m_pState->m_central_dir_offsets, mz_uint32, + file_index)); + mz_uint filename_len = MZ_READ_LE16(pHeader + MZ_ZIP_CDH_FILENAME_LEN_OFS); + const char *pFilename = + (const char *)pHeader + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE; + if (filename_len < name_len) continue; + if (comment_len) { + mz_uint file_extra_len = MZ_READ_LE16(pHeader + MZ_ZIP_CDH_EXTRA_LEN_OFS), + file_comment_len = + MZ_READ_LE16(pHeader + MZ_ZIP_CDH_COMMENT_LEN_OFS); + const char *pFile_comment = pFilename + filename_len + file_extra_len; + if ((file_comment_len != comment_len) || + (!mz_zip_reader_string_equal(pComment, pFile_comment, + file_comment_len, flags))) + continue; + } + if ((flags & MZ_ZIP_FLAG_IGNORE_PATH) && (filename_len)) { + int ofs = filename_len - 1; + do { + if ((pFilename[ofs] == '/') || (pFilename[ofs] == '\\') || + (pFilename[ofs] == ':')) + break; + } while (--ofs >= 0); + ofs++; + pFilename += ofs; + filename_len -= ofs; + } + if ((filename_len == name_len) && + (mz_zip_reader_string_equal(pName, pFilename, filename_len, flags))) + return file_index; + } + return -1; +} + +mz_bool mz_zip_reader_extract_to_mem_no_alloc(mz_zip_archive *pZip, + mz_uint file_index, void *pBuf, + size_t buf_size, mz_uint flags, + void *pUser_read_buf, + size_t user_read_buf_size) { + int status = TINFL_STATUS_DONE; + mz_uint64 needed_size, cur_file_ofs, comp_remaining, + out_buf_ofs = 0, read_buf_size, read_buf_ofs = 0, read_buf_avail; + mz_zip_archive_file_stat file_stat; + void *pRead_buf; + mz_uint32 + local_header_u32[(MZ_ZIP_LOCAL_DIR_HEADER_SIZE + sizeof(mz_uint32) - 1) / + sizeof(mz_uint32)]; + mz_uint8 *pLocal_header = (mz_uint8 *)local_header_u32; + tinfl_decompressor inflator; + + if ((buf_size) && (!pBuf)) return MZ_FALSE; + + if (!mz_zip_reader_file_stat(pZip, file_index, &file_stat)) return MZ_FALSE; + + // Empty file, or a directory (but not always a directory - I've seen odd zips + // with directories that have compressed data which inflates to 0 bytes) + if (!file_stat.m_comp_size) return MZ_TRUE; + + // Entry is a subdirectory (I've seen old zips with dir entries which have + // compressed deflate data which inflates to 0 bytes, but these entries claim + // to uncompress to 512 bytes in the headers). + // I'm torn how to handle this case - should it fail instead? + if (mz_zip_reader_is_file_a_directory(pZip, file_index)) return MZ_TRUE; + + // Encryption and patch files are not supported. + if (file_stat.m_bit_flag & (1 | 32)) return MZ_FALSE; + + // This function only supports stored and deflate. + if ((!(flags & MZ_ZIP_FLAG_COMPRESSED_DATA)) && (file_stat.m_method != 0) && + (file_stat.m_method != MZ_DEFLATED)) + return MZ_FALSE; + + // Ensure supplied output buffer is large enough. + needed_size = (flags & MZ_ZIP_FLAG_COMPRESSED_DATA) ? file_stat.m_comp_size + : file_stat.m_uncomp_size; + if (buf_size < needed_size) return MZ_FALSE; + + // Read and parse the local directory entry. + cur_file_ofs = file_stat.m_local_header_ofs; + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pLocal_header, + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) != + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) + return MZ_FALSE; + if (MZ_READ_LE32(pLocal_header) != MZ_ZIP_LOCAL_DIR_HEADER_SIG) + return MZ_FALSE; + + cur_file_ofs += MZ_ZIP_LOCAL_DIR_HEADER_SIZE + + MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_FILENAME_LEN_OFS) + + MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_EXTRA_LEN_OFS); + if ((cur_file_ofs + file_stat.m_comp_size) > pZip->m_archive_size) + return MZ_FALSE; + + if ((flags & MZ_ZIP_FLAG_COMPRESSED_DATA) || (!file_stat.m_method)) { + // The file is stored or the caller has requested the compressed data. + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pBuf, + (size_t)needed_size) != needed_size) + return MZ_FALSE; + return ((flags & MZ_ZIP_FLAG_COMPRESSED_DATA) != 0) || + (mz_crc32(MZ_CRC32_INIT, (const mz_uint8 *)pBuf, + (size_t)file_stat.m_uncomp_size) == file_stat.m_crc32); + } + + // Decompress the file either directly from memory or from a file input + // buffer. + tinfl_init(&inflator); + + if (pZip->m_pState->m_pMem) { + // Read directly from the archive in memory. + pRead_buf = (mz_uint8 *)pZip->m_pState->m_pMem + cur_file_ofs; + read_buf_size = read_buf_avail = file_stat.m_comp_size; + comp_remaining = 0; + } else if (pUser_read_buf) { + // Use a user provided read buffer. + if (!user_read_buf_size) return MZ_FALSE; + pRead_buf = (mz_uint8 *)pUser_read_buf; + read_buf_size = user_read_buf_size; + read_buf_avail = 0; + comp_remaining = file_stat.m_comp_size; + } else { + // Temporarily allocate a read buffer. + read_buf_size = + MZ_MIN(file_stat.m_comp_size, (mz_uint)MZ_ZIP_MAX_IO_BUF_SIZE); +#ifdef _MSC_VER + if (((0, sizeof(size_t) == sizeof(mz_uint32))) && + (read_buf_size > 0x7FFFFFFF)) +#else + if (((sizeof(size_t) == sizeof(mz_uint32))) && (read_buf_size > 0x7FFFFFFF)) +#endif + return MZ_FALSE; + if (NULL == (pRead_buf = pZip->m_pAlloc(pZip->m_pAlloc_opaque, 1, + (size_t)read_buf_size))) + return MZ_FALSE; + read_buf_avail = 0; + comp_remaining = file_stat.m_comp_size; + } + + do { + size_t in_buf_size, + out_buf_size = (size_t)(file_stat.m_uncomp_size - out_buf_ofs); + if ((!read_buf_avail) && (!pZip->m_pState->m_pMem)) { + read_buf_avail = MZ_MIN(read_buf_size, comp_remaining); + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pRead_buf, + (size_t)read_buf_avail) != read_buf_avail) { + status = TINFL_STATUS_FAILED; + break; + } + cur_file_ofs += read_buf_avail; + comp_remaining -= read_buf_avail; + read_buf_ofs = 0; + } + in_buf_size = (size_t)read_buf_avail; + status = tinfl_decompress( + &inflator, (mz_uint8 *)pRead_buf + read_buf_ofs, &in_buf_size, + (mz_uint8 *)pBuf, (mz_uint8 *)pBuf + out_buf_ofs, &out_buf_size, + TINFL_FLAG_USING_NON_WRAPPING_OUTPUT_BUF | + (comp_remaining ? TINFL_FLAG_HAS_MORE_INPUT : 0)); + read_buf_avail -= in_buf_size; + read_buf_ofs += in_buf_size; + out_buf_ofs += out_buf_size; + } while (status == TINFL_STATUS_NEEDS_MORE_INPUT); + + if (status == TINFL_STATUS_DONE) { + // Make sure the entire file was decompressed, and check its CRC. + if ((out_buf_ofs != file_stat.m_uncomp_size) || + (mz_crc32(MZ_CRC32_INIT, (const mz_uint8 *)pBuf, + (size_t)file_stat.m_uncomp_size) != file_stat.m_crc32)) + status = TINFL_STATUS_FAILED; + } + + if ((!pZip->m_pState->m_pMem) && (!pUser_read_buf)) + pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + + return status == TINFL_STATUS_DONE; +} + +mz_bool mz_zip_reader_extract_file_to_mem_no_alloc( + mz_zip_archive *pZip, const char *pFilename, void *pBuf, size_t buf_size, + mz_uint flags, void *pUser_read_buf, size_t user_read_buf_size) { + int file_index = mz_zip_reader_locate_file(pZip, pFilename, NULL, flags); + if (file_index < 0) return MZ_FALSE; + return mz_zip_reader_extract_to_mem_no_alloc(pZip, file_index, pBuf, buf_size, + flags, pUser_read_buf, + user_read_buf_size); +} + +mz_bool mz_zip_reader_extract_to_mem(mz_zip_archive *pZip, mz_uint file_index, + void *pBuf, size_t buf_size, + mz_uint flags) { + return mz_zip_reader_extract_to_mem_no_alloc(pZip, file_index, pBuf, buf_size, + flags, NULL, 0); +} + +mz_bool mz_zip_reader_extract_file_to_mem(mz_zip_archive *pZip, + const char *pFilename, void *pBuf, + size_t buf_size, mz_uint flags) { + return mz_zip_reader_extract_file_to_mem_no_alloc(pZip, pFilename, pBuf, + buf_size, flags, NULL, 0); +} + +void *mz_zip_reader_extract_to_heap(mz_zip_archive *pZip, mz_uint file_index, + size_t *pSize, mz_uint flags) { + mz_uint64 comp_size, uncomp_size, alloc_size; + const mz_uint8 *p = mz_zip_reader_get_cdh(pZip, file_index); + void *pBuf; + + if (pSize) *pSize = 0; + if (!p) return NULL; + + comp_size = MZ_READ_LE32(p + MZ_ZIP_CDH_COMPRESSED_SIZE_OFS); + uncomp_size = MZ_READ_LE32(p + MZ_ZIP_CDH_DECOMPRESSED_SIZE_OFS); + + alloc_size = (flags & MZ_ZIP_FLAG_COMPRESSED_DATA) ? comp_size : uncomp_size; +#ifdef _MSC_VER + if (((0, sizeof(size_t) == sizeof(mz_uint32))) && (alloc_size > 0x7FFFFFFF)) +#else + if (((sizeof(size_t) == sizeof(mz_uint32))) && (alloc_size > 0x7FFFFFFF)) +#endif + return NULL; + if (NULL == + (pBuf = pZip->m_pAlloc(pZip->m_pAlloc_opaque, 1, (size_t)alloc_size))) + return NULL; + + if (!mz_zip_reader_extract_to_mem(pZip, file_index, pBuf, (size_t)alloc_size, + flags)) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pBuf); + return NULL; + } + + if (pSize) *pSize = (size_t)alloc_size; + return pBuf; +} + +void *mz_zip_reader_extract_file_to_heap(mz_zip_archive *pZip, + const char *pFilename, size_t *pSize, + mz_uint flags) { + int file_index = mz_zip_reader_locate_file(pZip, pFilename, NULL, flags); + if (file_index < 0) { + if (pSize) *pSize = 0; + return MZ_FALSE; + } + return mz_zip_reader_extract_to_heap(pZip, file_index, pSize, flags); +} + +mz_bool mz_zip_reader_extract_to_callback(mz_zip_archive *pZip, + mz_uint file_index, + mz_file_write_func pCallback, + void *pOpaque, mz_uint flags) { + int status = TINFL_STATUS_DONE; + mz_uint file_crc32 = MZ_CRC32_INIT; + mz_uint64 read_buf_size, read_buf_ofs = 0, read_buf_avail, comp_remaining, + out_buf_ofs = 0, cur_file_ofs; + mz_zip_archive_file_stat file_stat; + void *pRead_buf = NULL; + void *pWrite_buf = NULL; + mz_uint32 + local_header_u32[(MZ_ZIP_LOCAL_DIR_HEADER_SIZE + sizeof(mz_uint32) - 1) / + sizeof(mz_uint32)]; + mz_uint8 *pLocal_header = (mz_uint8 *)local_header_u32; + + if (!mz_zip_reader_file_stat(pZip, file_index, &file_stat)) return MZ_FALSE; + + // Empty file, or a directory (but not always a directory - I've seen odd zips + // with directories that have compressed data which inflates to 0 bytes) + if (!file_stat.m_comp_size) return MZ_TRUE; + + // Entry is a subdirectory (I've seen old zips with dir entries which have + // compressed deflate data which inflates to 0 bytes, but these entries claim + // to uncompress to 512 bytes in the headers). + // I'm torn how to handle this case - should it fail instead? + if (mz_zip_reader_is_file_a_directory(pZip, file_index)) return MZ_TRUE; + + // Encryption and patch files are not supported. + if (file_stat.m_bit_flag & (1 | 32)) return MZ_FALSE; + + // This function only supports stored and deflate. + if ((!(flags & MZ_ZIP_FLAG_COMPRESSED_DATA)) && (file_stat.m_method != 0) && + (file_stat.m_method != MZ_DEFLATED)) + return MZ_FALSE; + + // Read and parse the local directory entry. + cur_file_ofs = file_stat.m_local_header_ofs; + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pLocal_header, + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) != + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) + return MZ_FALSE; + if (MZ_READ_LE32(pLocal_header) != MZ_ZIP_LOCAL_DIR_HEADER_SIG) + return MZ_FALSE; + + cur_file_ofs += MZ_ZIP_LOCAL_DIR_HEADER_SIZE + + MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_FILENAME_LEN_OFS) + + MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_EXTRA_LEN_OFS); + if ((cur_file_ofs + file_stat.m_comp_size) > pZip->m_archive_size) + return MZ_FALSE; + + // Decompress the file either directly from memory or from a file input + // buffer. + if (pZip->m_pState->m_pMem) { + pRead_buf = (mz_uint8 *)pZip->m_pState->m_pMem + cur_file_ofs; + read_buf_size = read_buf_avail = file_stat.m_comp_size; + comp_remaining = 0; + } else { + read_buf_size = + MZ_MIN(file_stat.m_comp_size, (mz_uint)MZ_ZIP_MAX_IO_BUF_SIZE); + if (NULL == (pRead_buf = pZip->m_pAlloc(pZip->m_pAlloc_opaque, 1, + (size_t)read_buf_size))) + return MZ_FALSE; + read_buf_avail = 0; + comp_remaining = file_stat.m_comp_size; + } + + if ((flags & MZ_ZIP_FLAG_COMPRESSED_DATA) || (!file_stat.m_method)) { + // The file is stored or the caller has requested the compressed data. + if (pZip->m_pState->m_pMem) { +#ifdef _MSC_VER + if (((0, sizeof(size_t) == sizeof(mz_uint32))) && + (file_stat.m_comp_size > 0xFFFFFFFF)) +#else + if (((sizeof(size_t) == sizeof(mz_uint32))) && + (file_stat.m_comp_size > 0xFFFFFFFF)) +#endif + return MZ_FALSE; + if (pCallback(pOpaque, out_buf_ofs, pRead_buf, + (size_t)file_stat.m_comp_size) != file_stat.m_comp_size) + status = TINFL_STATUS_FAILED; + else if (!(flags & MZ_ZIP_FLAG_COMPRESSED_DATA)) + file_crc32 = + (mz_uint32)mz_crc32(file_crc32, (const mz_uint8 *)pRead_buf, + (size_t)file_stat.m_comp_size); + cur_file_ofs += file_stat.m_comp_size; + out_buf_ofs += file_stat.m_comp_size; + comp_remaining = 0; + } else { + while (comp_remaining) { + read_buf_avail = MZ_MIN(read_buf_size, comp_remaining); + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pRead_buf, + (size_t)read_buf_avail) != read_buf_avail) { + status = TINFL_STATUS_FAILED; + break; + } + + if (!(flags & MZ_ZIP_FLAG_COMPRESSED_DATA)) + file_crc32 = (mz_uint32)mz_crc32( + file_crc32, (const mz_uint8 *)pRead_buf, (size_t)read_buf_avail); + + if (pCallback(pOpaque, out_buf_ofs, pRead_buf, + (size_t)read_buf_avail) != read_buf_avail) { + status = TINFL_STATUS_FAILED; + break; + } + cur_file_ofs += read_buf_avail; + out_buf_ofs += read_buf_avail; + comp_remaining -= read_buf_avail; + } + } + } else { + tinfl_decompressor inflator; + tinfl_init(&inflator); + + if (NULL == (pWrite_buf = pZip->m_pAlloc(pZip->m_pAlloc_opaque, 1, + TINFL_LZ_DICT_SIZE))) + status = TINFL_STATUS_FAILED; + else { + do { + mz_uint8 *pWrite_buf_cur = + (mz_uint8 *)pWrite_buf + (out_buf_ofs & (TINFL_LZ_DICT_SIZE - 1)); + size_t in_buf_size, + out_buf_size = + TINFL_LZ_DICT_SIZE - (out_buf_ofs & (TINFL_LZ_DICT_SIZE - 1)); + if ((!read_buf_avail) && (!pZip->m_pState->m_pMem)) { + read_buf_avail = MZ_MIN(read_buf_size, comp_remaining); + if (pZip->m_pRead(pZip->m_pIO_opaque, cur_file_ofs, pRead_buf, + (size_t)read_buf_avail) != read_buf_avail) { + status = TINFL_STATUS_FAILED; + break; + } + cur_file_ofs += read_buf_avail; + comp_remaining -= read_buf_avail; + read_buf_ofs = 0; + } + + in_buf_size = (size_t)read_buf_avail; + status = tinfl_decompress( + &inflator, (const mz_uint8 *)pRead_buf + read_buf_ofs, &in_buf_size, + (mz_uint8 *)pWrite_buf, pWrite_buf_cur, &out_buf_size, + comp_remaining ? TINFL_FLAG_HAS_MORE_INPUT : 0); + read_buf_avail -= in_buf_size; + read_buf_ofs += in_buf_size; + + if (out_buf_size) { + if (pCallback(pOpaque, out_buf_ofs, pWrite_buf_cur, out_buf_size) != + out_buf_size) { + status = TINFL_STATUS_FAILED; + break; + } + file_crc32 = + (mz_uint32)mz_crc32(file_crc32, pWrite_buf_cur, out_buf_size); + if ((out_buf_ofs += out_buf_size) > file_stat.m_uncomp_size) { + status = TINFL_STATUS_FAILED; + break; + } + } + } while ((status == TINFL_STATUS_NEEDS_MORE_INPUT) || + (status == TINFL_STATUS_HAS_MORE_OUTPUT)); + } + } + + if ((status == TINFL_STATUS_DONE) && + (!(flags & MZ_ZIP_FLAG_COMPRESSED_DATA))) { + // Make sure the entire file was decompressed, and check its CRC. + if ((out_buf_ofs != file_stat.m_uncomp_size) || + (file_crc32 != file_stat.m_crc32)) + status = TINFL_STATUS_FAILED; + } + + if (!pZip->m_pState->m_pMem) pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + if (pWrite_buf) pZip->m_pFree(pZip->m_pAlloc_opaque, pWrite_buf); + + return status == TINFL_STATUS_DONE; +} + +mz_bool mz_zip_reader_extract_file_to_callback(mz_zip_archive *pZip, + const char *pFilename, + mz_file_write_func pCallback, + void *pOpaque, mz_uint flags) { + int file_index = mz_zip_reader_locate_file(pZip, pFilename, NULL, flags); + if (file_index < 0) return MZ_FALSE; + return mz_zip_reader_extract_to_callback(pZip, file_index, pCallback, pOpaque, + flags); +} + +#ifndef MINIZ_NO_STDIO +static size_t mz_zip_file_write_callback(void *pOpaque, mz_uint64 ofs, + const void *pBuf, size_t n) { + (void)ofs; + return MZ_FWRITE(pBuf, 1, n, (MZ_FILE *)pOpaque); +} + +mz_bool mz_zip_reader_extract_to_file(mz_zip_archive *pZip, mz_uint file_index, + const char *pDst_filename, + mz_uint flags) { + mz_bool status; + mz_zip_archive_file_stat file_stat; + MZ_FILE *pFile; + if (!mz_zip_reader_file_stat(pZip, file_index, &file_stat)) return MZ_FALSE; + pFile = MZ_FOPEN(pDst_filename, "wb"); + if (!pFile) return MZ_FALSE; + status = mz_zip_reader_extract_to_callback( + pZip, file_index, mz_zip_file_write_callback, pFile, flags); + if (MZ_FCLOSE(pFile) == EOF) return MZ_FALSE; +#ifndef MINIZ_NO_TIME + if (status) + mz_zip_set_file_times(pDst_filename, file_stat.m_time, file_stat.m_time); +#endif + return status; +} +#endif // #ifndef MINIZ_NO_STDIO + +mz_bool mz_zip_reader_end(mz_zip_archive *pZip) { + if ((!pZip) || (!pZip->m_pState) || (!pZip->m_pAlloc) || (!pZip->m_pFree) || + (pZip->m_zip_mode != MZ_ZIP_MODE_READING)) + return MZ_FALSE; + + if (pZip->m_pState) { + mz_zip_internal_state *pState = pZip->m_pState; + pZip->m_pState = NULL; + mz_zip_array_clear(pZip, &pState->m_central_dir); + mz_zip_array_clear(pZip, &pState->m_central_dir_offsets); + mz_zip_array_clear(pZip, &pState->m_sorted_central_dir_offsets); + +#ifndef MINIZ_NO_STDIO + if (pState->m_pFile) { + MZ_FCLOSE(pState->m_pFile); + pState->m_pFile = NULL; + } +#endif // #ifndef MINIZ_NO_STDIO + + pZip->m_pFree(pZip->m_pAlloc_opaque, pState); + } + pZip->m_zip_mode = MZ_ZIP_MODE_INVALID; + + return MZ_TRUE; +} + +#ifndef MINIZ_NO_STDIO +mz_bool mz_zip_reader_extract_file_to_file(mz_zip_archive *pZip, + const char *pArchive_filename, + const char *pDst_filename, + mz_uint flags) { + int file_index = + mz_zip_reader_locate_file(pZip, pArchive_filename, NULL, flags); + if (file_index < 0) return MZ_FALSE; + return mz_zip_reader_extract_to_file(pZip, file_index, pDst_filename, flags); +} +#endif + +// ------------------- .ZIP archive writing + +#ifndef MINIZ_NO_ARCHIVE_WRITING_APIS + +static void mz_write_le16(mz_uint8 *p, mz_uint16 v) { + p[0] = (mz_uint8)v; + p[1] = (mz_uint8)(v >> 8); +} +static void mz_write_le32(mz_uint8 *p, mz_uint32 v) { + p[0] = (mz_uint8)v; + p[1] = (mz_uint8)(v >> 8); + p[2] = (mz_uint8)(v >> 16); + p[3] = (mz_uint8)(v >> 24); +} +#define MZ_WRITE_LE16(p, v) mz_write_le16((mz_uint8 *)(p), (mz_uint16)(v)) +#define MZ_WRITE_LE32(p, v) mz_write_le32((mz_uint8 *)(p), (mz_uint32)(v)) + +mz_bool mz_zip_writer_init(mz_zip_archive *pZip, mz_uint64 existing_size) { + if ((!pZip) || (pZip->m_pState) || (!pZip->m_pWrite) || + (pZip->m_zip_mode != MZ_ZIP_MODE_INVALID)) + return MZ_FALSE; + + if (pZip->m_file_offset_alignment) { + // Ensure user specified file offset alignment is a power of 2. + if (pZip->m_file_offset_alignment & (pZip->m_file_offset_alignment - 1)) + return MZ_FALSE; + } + + if (!pZip->m_pAlloc) pZip->m_pAlloc = def_alloc_func; + if (!pZip->m_pFree) pZip->m_pFree = def_free_func; + if (!pZip->m_pRealloc) pZip->m_pRealloc = def_realloc_func; + + pZip->m_zip_mode = MZ_ZIP_MODE_WRITING; + pZip->m_archive_size = existing_size; + pZip->m_central_directory_file_ofs = 0; + pZip->m_total_files = 0; + + if (NULL == (pZip->m_pState = (mz_zip_internal_state *)pZip->m_pAlloc( + pZip->m_pAlloc_opaque, 1, sizeof(mz_zip_internal_state)))) + return MZ_FALSE; + memset(pZip->m_pState, 0, sizeof(mz_zip_internal_state)); + MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(&pZip->m_pState->m_central_dir, + sizeof(mz_uint8)); + MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(&pZip->m_pState->m_central_dir_offsets, + sizeof(mz_uint32)); + MZ_ZIP_ARRAY_SET_ELEMENT_SIZE(&pZip->m_pState->m_sorted_central_dir_offsets, + sizeof(mz_uint32)); + return MZ_TRUE; +} + +static size_t mz_zip_heap_write_func(void *pOpaque, mz_uint64 file_ofs, + const void *pBuf, size_t n) { + mz_zip_archive *pZip = (mz_zip_archive *)pOpaque; + mz_zip_internal_state *pState = pZip->m_pState; + mz_uint64 new_size = MZ_MAX(file_ofs + n, pState->m_mem_size); +#ifdef _MSC_VER + if ((!n) || + ((0, sizeof(size_t) == sizeof(mz_uint32)) && (new_size > 0x7FFFFFFF))) +#else + if ((!n) || + ((sizeof(size_t) == sizeof(mz_uint32)) && (new_size > 0x7FFFFFFF))) +#endif + return 0; + if (new_size > pState->m_mem_capacity) { + void *pNew_block; + size_t new_capacity = MZ_MAX(64, pState->m_mem_capacity); + while (new_capacity < new_size) new_capacity *= 2; + if (NULL == (pNew_block = pZip->m_pRealloc( + pZip->m_pAlloc_opaque, pState->m_pMem, 1, new_capacity))) + return 0; + pState->m_pMem = pNew_block; + pState->m_mem_capacity = new_capacity; + } + memcpy((mz_uint8 *)pState->m_pMem + file_ofs, pBuf, n); + pState->m_mem_size = (size_t)new_size; + return n; +} + +mz_bool mz_zip_writer_init_heap(mz_zip_archive *pZip, + size_t size_to_reserve_at_beginning, + size_t initial_allocation_size) { + pZip->m_pWrite = mz_zip_heap_write_func; + pZip->m_pIO_opaque = pZip; + if (!mz_zip_writer_init(pZip, size_to_reserve_at_beginning)) return MZ_FALSE; + if (0 != (initial_allocation_size = MZ_MAX(initial_allocation_size, + size_to_reserve_at_beginning))) { + if (NULL == (pZip->m_pState->m_pMem = pZip->m_pAlloc( + pZip->m_pAlloc_opaque, 1, initial_allocation_size))) { + mz_zip_writer_end(pZip); + return MZ_FALSE; + } + pZip->m_pState->m_mem_capacity = initial_allocation_size; + } + return MZ_TRUE; +} + +#ifndef MINIZ_NO_STDIO +static size_t mz_zip_file_write_func(void *pOpaque, mz_uint64 file_ofs, + const void *pBuf, size_t n) { + mz_zip_archive *pZip = (mz_zip_archive *)pOpaque; + mz_int64 cur_ofs = MZ_FTELL64(pZip->m_pState->m_pFile); + if (((mz_int64)file_ofs < 0) || + (((cur_ofs != (mz_int64)file_ofs)) && + (MZ_FSEEK64(pZip->m_pState->m_pFile, (mz_int64)file_ofs, SEEK_SET)))) + return 0; + return MZ_FWRITE(pBuf, 1, n, pZip->m_pState->m_pFile); +} + +mz_bool mz_zip_writer_init_file(mz_zip_archive *pZip, const char *pFilename, + mz_uint64 size_to_reserve_at_beginning) { + MZ_FILE *pFile; + pZip->m_pWrite = mz_zip_file_write_func; + pZip->m_pIO_opaque = pZip; + if (!mz_zip_writer_init(pZip, size_to_reserve_at_beginning)) return MZ_FALSE; + if (NULL == (pFile = MZ_FOPEN(pFilename, "wb"))) { + mz_zip_writer_end(pZip); + return MZ_FALSE; + } + pZip->m_pState->m_pFile = pFile; + if (size_to_reserve_at_beginning) { + mz_uint64 cur_ofs = 0; + char buf[4096]; + MZ_CLEAR_OBJ(buf); + do { + size_t n = (size_t)MZ_MIN(sizeof(buf), size_to_reserve_at_beginning); + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_ofs, buf, n) != n) { + mz_zip_writer_end(pZip); + return MZ_FALSE; + } + cur_ofs += n; + size_to_reserve_at_beginning -= n; + } while (size_to_reserve_at_beginning); + } + return MZ_TRUE; +} +#endif // #ifndef MINIZ_NO_STDIO + +mz_bool mz_zip_writer_init_from_reader(mz_zip_archive *pZip, + const char *pFilename) { + mz_zip_internal_state *pState; + if ((!pZip) || (!pZip->m_pState) || (pZip->m_zip_mode != MZ_ZIP_MODE_READING)) + return MZ_FALSE; + // No sense in trying to write to an archive that's already at the support max + // size + if ((pZip->m_total_files == 0xFFFF) || + ((pZip->m_archive_size + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) > 0xFFFFFFFF)) + return MZ_FALSE; + + pState = pZip->m_pState; + + if (pState->m_pFile) { +#ifdef MINIZ_NO_STDIO + pFilename; + return MZ_FALSE; +#else + // Archive is being read from stdio - try to reopen as writable. + if (pZip->m_pIO_opaque != pZip) return MZ_FALSE; + if (!pFilename) return MZ_FALSE; + pZip->m_pWrite = mz_zip_file_write_func; + if (NULL == + (pState->m_pFile = MZ_FREOPEN(pFilename, "r+b", pState->m_pFile))) { + // The mz_zip_archive is now in a bogus state because pState->m_pFile is + // NULL, so just close it. + mz_zip_reader_end(pZip); + return MZ_FALSE; + } +#endif // #ifdef MINIZ_NO_STDIO + } else if (pState->m_pMem) { + // Archive lives in a memory block. Assume it's from the heap that we can + // resize using the realloc callback. + if (pZip->m_pIO_opaque != pZip) return MZ_FALSE; + pState->m_mem_capacity = pState->m_mem_size; + pZip->m_pWrite = mz_zip_heap_write_func; + } + // Archive is being read via a user provided read function - make sure the + // user has specified a write function too. + else if (!pZip->m_pWrite) + return MZ_FALSE; + + // Start writing new files at the archive's current central directory + // location. + pZip->m_archive_size = pZip->m_central_directory_file_ofs; + pZip->m_zip_mode = MZ_ZIP_MODE_WRITING; + pZip->m_central_directory_file_ofs = 0; + + return MZ_TRUE; +} + +mz_bool mz_zip_writer_add_mem(mz_zip_archive *pZip, const char *pArchive_name, + const void *pBuf, size_t buf_size, + mz_uint level_and_flags) { + return mz_zip_writer_add_mem_ex(pZip, pArchive_name, pBuf, buf_size, NULL, 0, + level_and_flags, 0, 0); +} + +typedef struct { + mz_zip_archive *m_pZip; + mz_uint64 m_cur_archive_file_ofs; + mz_uint64 m_comp_size; +} mz_zip_writer_add_state; + +static mz_bool mz_zip_writer_add_put_buf_callback(const void *pBuf, int len, + void *pUser) { + mz_zip_writer_add_state *pState = (mz_zip_writer_add_state *)pUser; + if ((int)pState->m_pZip->m_pWrite(pState->m_pZip->m_pIO_opaque, + pState->m_cur_archive_file_ofs, pBuf, + len) != len) + return MZ_FALSE; + pState->m_cur_archive_file_ofs += len; + pState->m_comp_size += len; + return MZ_TRUE; +} + +static mz_bool mz_zip_writer_create_local_dir_header( + mz_zip_archive *pZip, mz_uint8 *pDst, mz_uint16 filename_size, + mz_uint16 extra_size, mz_uint64 uncomp_size, mz_uint64 comp_size, + mz_uint32 uncomp_crc32, mz_uint16 method, mz_uint16 bit_flags, + mz_uint16 dos_time, mz_uint16 dos_date) { + (void)pZip; + memset(pDst, 0, MZ_ZIP_LOCAL_DIR_HEADER_SIZE); + MZ_WRITE_LE32(pDst + MZ_ZIP_LDH_SIG_OFS, MZ_ZIP_LOCAL_DIR_HEADER_SIG); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_VERSION_NEEDED_OFS, method ? 20 : 0); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_BIT_FLAG_OFS, bit_flags); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_METHOD_OFS, method); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_FILE_TIME_OFS, dos_time); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_FILE_DATE_OFS, dos_date); + MZ_WRITE_LE32(pDst + MZ_ZIP_LDH_CRC32_OFS, uncomp_crc32); + MZ_WRITE_LE32(pDst + MZ_ZIP_LDH_COMPRESSED_SIZE_OFS, comp_size); + MZ_WRITE_LE32(pDst + MZ_ZIP_LDH_DECOMPRESSED_SIZE_OFS, uncomp_size); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_FILENAME_LEN_OFS, filename_size); + MZ_WRITE_LE16(pDst + MZ_ZIP_LDH_EXTRA_LEN_OFS, extra_size); + return MZ_TRUE; +} + +static mz_bool mz_zip_writer_create_central_dir_header( + mz_zip_archive *pZip, mz_uint8 *pDst, mz_uint16 filename_size, + mz_uint16 extra_size, mz_uint16 comment_size, mz_uint64 uncomp_size, + mz_uint64 comp_size, mz_uint32 uncomp_crc32, mz_uint16 method, + mz_uint16 bit_flags, mz_uint16 dos_time, mz_uint16 dos_date, + mz_uint64 local_header_ofs, mz_uint32 ext_attributes) { + (void)pZip; + memset(pDst, 0, MZ_ZIP_CENTRAL_DIR_HEADER_SIZE); + MZ_WRITE_LE32(pDst + MZ_ZIP_CDH_SIG_OFS, MZ_ZIP_CENTRAL_DIR_HEADER_SIG); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_VERSION_NEEDED_OFS, method ? 20 : 0); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_BIT_FLAG_OFS, bit_flags); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_METHOD_OFS, method); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_FILE_TIME_OFS, dos_time); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_FILE_DATE_OFS, dos_date); + MZ_WRITE_LE32(pDst + MZ_ZIP_CDH_CRC32_OFS, uncomp_crc32); + MZ_WRITE_LE32(pDst + MZ_ZIP_CDH_COMPRESSED_SIZE_OFS, comp_size); + MZ_WRITE_LE32(pDst + MZ_ZIP_CDH_DECOMPRESSED_SIZE_OFS, uncomp_size); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_FILENAME_LEN_OFS, filename_size); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_EXTRA_LEN_OFS, extra_size); + MZ_WRITE_LE16(pDst + MZ_ZIP_CDH_COMMENT_LEN_OFS, comment_size); + MZ_WRITE_LE32(pDst + MZ_ZIP_CDH_EXTERNAL_ATTR_OFS, ext_attributes); + MZ_WRITE_LE32(pDst + MZ_ZIP_CDH_LOCAL_HEADER_OFS, local_header_ofs); + return MZ_TRUE; +} + +static mz_bool mz_zip_writer_add_to_central_dir( + mz_zip_archive *pZip, const char *pFilename, mz_uint16 filename_size, + const void *pExtra, mz_uint16 extra_size, const void *pComment, + mz_uint16 comment_size, mz_uint64 uncomp_size, mz_uint64 comp_size, + mz_uint32 uncomp_crc32, mz_uint16 method, mz_uint16 bit_flags, + mz_uint16 dos_time, mz_uint16 dos_date, mz_uint64 local_header_ofs, + mz_uint32 ext_attributes) { + mz_zip_internal_state *pState = pZip->m_pState; + mz_uint32 central_dir_ofs = (mz_uint32)pState->m_central_dir.m_size; + size_t orig_central_dir_size = pState->m_central_dir.m_size; + mz_uint8 central_dir_header[MZ_ZIP_CENTRAL_DIR_HEADER_SIZE]; + + // No zip64 support yet + if ((local_header_ofs > 0xFFFFFFFF) || + (((mz_uint64)pState->m_central_dir.m_size + + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + filename_size + extra_size + + comment_size) > 0xFFFFFFFF)) + return MZ_FALSE; + + if (!mz_zip_writer_create_central_dir_header( + pZip, central_dir_header, filename_size, extra_size, comment_size, + uncomp_size, comp_size, uncomp_crc32, method, bit_flags, dos_time, + dos_date, local_header_ofs, ext_attributes)) + return MZ_FALSE; + + if ((!mz_zip_array_push_back(pZip, &pState->m_central_dir, central_dir_header, + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE)) || + (!mz_zip_array_push_back(pZip, &pState->m_central_dir, pFilename, + filename_size)) || + (!mz_zip_array_push_back(pZip, &pState->m_central_dir, pExtra, + extra_size)) || + (!mz_zip_array_push_back(pZip, &pState->m_central_dir, pComment, + comment_size)) || + (!mz_zip_array_push_back(pZip, &pState->m_central_dir_offsets, + ¢ral_dir_ofs, 1))) { + // Try to push the central directory array back into its original state. + mz_zip_array_resize(pZip, &pState->m_central_dir, orig_central_dir_size, + MZ_FALSE); + return MZ_FALSE; + } + + return MZ_TRUE; +} + +static mz_bool mz_zip_writer_validate_archive_name(const char *pArchive_name) { + // Basic ZIP archive filename validity checks: Valid filenames cannot start + // with a forward slash, cannot contain a drive letter, and cannot use + // DOS-style backward slashes. + if (*pArchive_name == '/') return MZ_FALSE; + while (*pArchive_name) { + if ((*pArchive_name == '\\') || (*pArchive_name == ':')) return MZ_FALSE; + pArchive_name++; + } + return MZ_TRUE; +} + +static mz_uint mz_zip_writer_compute_padding_needed_for_file_alignment( + mz_zip_archive *pZip) { + mz_uint32 n; + if (!pZip->m_file_offset_alignment) return 0; + n = (mz_uint32)(pZip->m_archive_size & (pZip->m_file_offset_alignment - 1)); + return (pZip->m_file_offset_alignment - n) & + (pZip->m_file_offset_alignment - 1); +} + +static mz_bool mz_zip_writer_write_zeros(mz_zip_archive *pZip, + mz_uint64 cur_file_ofs, mz_uint32 n) { + char buf[4096]; + memset(buf, 0, MZ_MIN(sizeof(buf), n)); + while (n) { + mz_uint32 s = MZ_MIN(sizeof(buf), n); + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_file_ofs, buf, s) != s) + return MZ_FALSE; + cur_file_ofs += s; + n -= s; + } + return MZ_TRUE; +} + +mz_bool mz_zip_writer_add_mem_ex(mz_zip_archive *pZip, + const char *pArchive_name, const void *pBuf, + size_t buf_size, const void *pComment, + mz_uint16 comment_size, + mz_uint level_and_flags, mz_uint64 uncomp_size, + mz_uint32 uncomp_crc32) { + mz_uint16 method = 0, dos_time = 0, dos_date = 0; + mz_uint level, ext_attributes = 0, num_alignment_padding_bytes; + mz_uint64 local_dir_header_ofs = pZip->m_archive_size, + cur_archive_file_ofs = pZip->m_archive_size, comp_size = 0; + size_t archive_name_size; + mz_uint8 local_dir_header[MZ_ZIP_LOCAL_DIR_HEADER_SIZE]; + tdefl_compressor *pComp = NULL; + mz_bool store_data_uncompressed; + mz_zip_internal_state *pState; + + if ((int)level_and_flags < 0) level_and_flags = MZ_DEFAULT_LEVEL; + level = level_and_flags & 0xF; + store_data_uncompressed = + ((!level) || (level_and_flags & MZ_ZIP_FLAG_COMPRESSED_DATA)); + + if ((!pZip) || (!pZip->m_pState) || + (pZip->m_zip_mode != MZ_ZIP_MODE_WRITING) || ((buf_size) && (!pBuf)) || + (!pArchive_name) || ((comment_size) && (!pComment)) || + (pZip->m_total_files == 0xFFFF) || (level > MZ_UBER_COMPRESSION)) + return MZ_FALSE; + + pState = pZip->m_pState; + + if ((!(level_and_flags & MZ_ZIP_FLAG_COMPRESSED_DATA)) && (uncomp_size)) + return MZ_FALSE; + // No zip64 support yet + if ((buf_size > 0xFFFFFFFF) || (uncomp_size > 0xFFFFFFFF)) return MZ_FALSE; + if (!mz_zip_writer_validate_archive_name(pArchive_name)) return MZ_FALSE; + +#ifndef MINIZ_NO_TIME + { + time_t cur_time; + time(&cur_time); + mz_zip_time_to_dos_time(cur_time, &dos_time, &dos_date); + } +#endif // #ifndef MINIZ_NO_TIME + + archive_name_size = strlen(pArchive_name); + if (archive_name_size > 0xFFFF) return MZ_FALSE; + + num_alignment_padding_bytes = + mz_zip_writer_compute_padding_needed_for_file_alignment(pZip); + + // no zip64 support yet + if ((pZip->m_total_files == 0xFFFF) || + ((pZip->m_archive_size + num_alignment_padding_bytes + + MZ_ZIP_LOCAL_DIR_HEADER_SIZE + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + + comment_size + archive_name_size) > 0xFFFFFFFF)) + return MZ_FALSE; + + if ((archive_name_size) && (pArchive_name[archive_name_size - 1] == '/')) { + // Set DOS Subdirectory attribute bit. + ext_attributes |= 0x10; + // Subdirectories cannot contain data. + if ((buf_size) || (uncomp_size)) return MZ_FALSE; + } + + // Try to do any allocations before writing to the archive, so if an + // allocation fails the file remains unmodified. (A good idea if we're doing + // an in-place modification.) + if ((!mz_zip_array_ensure_room( + pZip, &pState->m_central_dir, + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + archive_name_size + comment_size)) || + (!mz_zip_array_ensure_room(pZip, &pState->m_central_dir_offsets, 1))) + return MZ_FALSE; + + if ((!store_data_uncompressed) && (buf_size)) { + if (NULL == (pComp = (tdefl_compressor *)pZip->m_pAlloc( + pZip->m_pAlloc_opaque, 1, sizeof(tdefl_compressor)))) + return MZ_FALSE; + } + + if (!mz_zip_writer_write_zeros( + pZip, cur_archive_file_ofs, + num_alignment_padding_bytes + sizeof(local_dir_header))) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + return MZ_FALSE; + } + local_dir_header_ofs += num_alignment_padding_bytes; + if (pZip->m_file_offset_alignment) { + MZ_ASSERT((local_dir_header_ofs & (pZip->m_file_offset_alignment - 1)) == + 0); + } + cur_archive_file_ofs += + num_alignment_padding_bytes + sizeof(local_dir_header); + + MZ_CLEAR_OBJ(local_dir_header); + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_archive_file_ofs, pArchive_name, + archive_name_size) != archive_name_size) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + return MZ_FALSE; + } + cur_archive_file_ofs += archive_name_size; + + if (!(level_and_flags & MZ_ZIP_FLAG_COMPRESSED_DATA)) { + uncomp_crc32 = + (mz_uint32)mz_crc32(MZ_CRC32_INIT, (const mz_uint8 *)pBuf, buf_size); + uncomp_size = buf_size; + if (uncomp_size <= 3) { + level = 0; + store_data_uncompressed = MZ_TRUE; + } + } + + if (store_data_uncompressed) { + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_archive_file_ofs, pBuf, + buf_size) != buf_size) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + return MZ_FALSE; + } + + cur_archive_file_ofs += buf_size; + comp_size = buf_size; + + if (level_and_flags & MZ_ZIP_FLAG_COMPRESSED_DATA) method = MZ_DEFLATED; + } else if (buf_size) { + mz_zip_writer_add_state state; + + state.m_pZip = pZip; + state.m_cur_archive_file_ofs = cur_archive_file_ofs; + state.m_comp_size = 0; + + if ((tdefl_init(pComp, mz_zip_writer_add_put_buf_callback, &state, + tdefl_create_comp_flags_from_zip_params( + level, -15, MZ_DEFAULT_STRATEGY)) != + TDEFL_STATUS_OKAY) || + (tdefl_compress_buffer(pComp, pBuf, buf_size, TDEFL_FINISH) != + TDEFL_STATUS_DONE)) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + return MZ_FALSE; + } + + comp_size = state.m_comp_size; + cur_archive_file_ofs = state.m_cur_archive_file_ofs; + + method = MZ_DEFLATED; + } + + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + pComp = NULL; + + // no zip64 support yet + if ((comp_size > 0xFFFFFFFF) || (cur_archive_file_ofs > 0xFFFFFFFF)) + return MZ_FALSE; + + if (!mz_zip_writer_create_local_dir_header( + pZip, local_dir_header, (mz_uint16)archive_name_size, 0, uncomp_size, + comp_size, uncomp_crc32, method, 0, dos_time, dos_date)) + return MZ_FALSE; + + if (pZip->m_pWrite(pZip->m_pIO_opaque, local_dir_header_ofs, local_dir_header, + sizeof(local_dir_header)) != sizeof(local_dir_header)) + return MZ_FALSE; + + if (!mz_zip_writer_add_to_central_dir( + pZip, pArchive_name, (mz_uint16)archive_name_size, NULL, 0, pComment, + comment_size, uncomp_size, comp_size, uncomp_crc32, method, 0, + dos_time, dos_date, local_dir_header_ofs, ext_attributes)) + return MZ_FALSE; + + pZip->m_total_files++; + pZip->m_archive_size = cur_archive_file_ofs; + + return MZ_TRUE; +} + +#ifndef MINIZ_NO_STDIO +mz_bool mz_zip_writer_add_file(mz_zip_archive *pZip, const char *pArchive_name, + const char *pSrc_filename, const void *pComment, + mz_uint16 comment_size, + mz_uint level_and_flags) { + mz_uint uncomp_crc32 = MZ_CRC32_INIT, level, num_alignment_padding_bytes; + mz_uint16 method = 0, dos_time = 0, dos_date = 0, ext_attributes = 0; + mz_uint64 local_dir_header_ofs = pZip->m_archive_size, + cur_archive_file_ofs = pZip->m_archive_size, uncomp_size = 0, + comp_size = 0; + size_t archive_name_size; + mz_uint8 local_dir_header[MZ_ZIP_LOCAL_DIR_HEADER_SIZE]; + MZ_FILE *pSrc_file = NULL; + + if ((int)level_and_flags < 0) level_and_flags = MZ_DEFAULT_LEVEL; + level = level_and_flags & 0xF; + + if ((!pZip) || (!pZip->m_pState) || + (pZip->m_zip_mode != MZ_ZIP_MODE_WRITING) || (!pArchive_name) || + ((comment_size) && (!pComment)) || (level > MZ_UBER_COMPRESSION)) + return MZ_FALSE; + if (level_and_flags & MZ_ZIP_FLAG_COMPRESSED_DATA) return MZ_FALSE; + if (!mz_zip_writer_validate_archive_name(pArchive_name)) return MZ_FALSE; + + archive_name_size = strlen(pArchive_name); + if (archive_name_size > 0xFFFF) return MZ_FALSE; + + num_alignment_padding_bytes = + mz_zip_writer_compute_padding_needed_for_file_alignment(pZip); + + // no zip64 support yet + if ((pZip->m_total_files == 0xFFFF) || + ((pZip->m_archive_size + num_alignment_padding_bytes + + MZ_ZIP_LOCAL_DIR_HEADER_SIZE + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE + + comment_size + archive_name_size) > 0xFFFFFFFF)) + return MZ_FALSE; + + if (!mz_zip_get_file_modified_time(pSrc_filename, &dos_time, &dos_date)) + return MZ_FALSE; + + pSrc_file = MZ_FOPEN(pSrc_filename, "rb"); + if (!pSrc_file) return MZ_FALSE; + MZ_FSEEK64(pSrc_file, 0, SEEK_END); + uncomp_size = MZ_FTELL64(pSrc_file); + MZ_FSEEK64(pSrc_file, 0, SEEK_SET); + + if (uncomp_size > 0xFFFFFFFF) { + // No zip64 support yet + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + if (uncomp_size <= 3) level = 0; + + if (!mz_zip_writer_write_zeros( + pZip, cur_archive_file_ofs, + num_alignment_padding_bytes + sizeof(local_dir_header))) { + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + local_dir_header_ofs += num_alignment_padding_bytes; + if (pZip->m_file_offset_alignment) { + MZ_ASSERT((local_dir_header_ofs & (pZip->m_file_offset_alignment - 1)) == + 0); + } + cur_archive_file_ofs += + num_alignment_padding_bytes + sizeof(local_dir_header); + + MZ_CLEAR_OBJ(local_dir_header); + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_archive_file_ofs, pArchive_name, + archive_name_size) != archive_name_size) { + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + cur_archive_file_ofs += archive_name_size; + + if (uncomp_size) { + mz_uint64 uncomp_remaining = uncomp_size; + void *pRead_buf = + pZip->m_pAlloc(pZip->m_pAlloc_opaque, 1, MZ_ZIP_MAX_IO_BUF_SIZE); + if (!pRead_buf) { + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + + if (!level) { + while (uncomp_remaining) { + mz_uint n = + (mz_uint)MZ_MIN((mz_uint)MZ_ZIP_MAX_IO_BUF_SIZE, uncomp_remaining); + if ((MZ_FREAD(pRead_buf, 1, n, pSrc_file) != n) || + (pZip->m_pWrite(pZip->m_pIO_opaque, cur_archive_file_ofs, pRead_buf, + n) != n)) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + uncomp_crc32 = + (mz_uint32)mz_crc32(uncomp_crc32, (const mz_uint8 *)pRead_buf, n); + uncomp_remaining -= n; + cur_archive_file_ofs += n; + } + comp_size = uncomp_size; + } else { + mz_bool result = MZ_FALSE; + mz_zip_writer_add_state state; + tdefl_compressor *pComp = (tdefl_compressor *)pZip->m_pAlloc( + pZip->m_pAlloc_opaque, 1, sizeof(tdefl_compressor)); + if (!pComp) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + + state.m_pZip = pZip; + state.m_cur_archive_file_ofs = cur_archive_file_ofs; + state.m_comp_size = 0; + + if (tdefl_init(pComp, mz_zip_writer_add_put_buf_callback, &state, + tdefl_create_comp_flags_from_zip_params( + level, -15, MZ_DEFAULT_STRATEGY)) != + TDEFL_STATUS_OKAY) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + + for (;;) { + size_t in_buf_size = (mz_uint32)MZ_MIN(uncomp_remaining, + (mz_uint)MZ_ZIP_MAX_IO_BUF_SIZE); + tdefl_status status; + + if (MZ_FREAD(pRead_buf, 1, in_buf_size, pSrc_file) != in_buf_size) + break; + + uncomp_crc32 = (mz_uint32)mz_crc32( + uncomp_crc32, (const mz_uint8 *)pRead_buf, in_buf_size); + uncomp_remaining -= in_buf_size; + + status = tdefl_compress_buffer( + pComp, pRead_buf, in_buf_size, + uncomp_remaining ? TDEFL_NO_FLUSH : TDEFL_FINISH); + if (status == TDEFL_STATUS_DONE) { + result = MZ_TRUE; + break; + } else if (status != TDEFL_STATUS_OKAY) + break; + } + + pZip->m_pFree(pZip->m_pAlloc_opaque, pComp); + + if (!result) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + MZ_FCLOSE(pSrc_file); + return MZ_FALSE; + } + + comp_size = state.m_comp_size; + cur_archive_file_ofs = state.m_cur_archive_file_ofs; + + method = MZ_DEFLATED; + } + + pZip->m_pFree(pZip->m_pAlloc_opaque, pRead_buf); + } + + MZ_FCLOSE(pSrc_file); + pSrc_file = NULL; + + // no zip64 support yet + if ((comp_size > 0xFFFFFFFF) || (cur_archive_file_ofs > 0xFFFFFFFF)) + return MZ_FALSE; + + if (!mz_zip_writer_create_local_dir_header( + pZip, local_dir_header, (mz_uint16)archive_name_size, 0, uncomp_size, + comp_size, uncomp_crc32, method, 0, dos_time, dos_date)) + return MZ_FALSE; + + if (pZip->m_pWrite(pZip->m_pIO_opaque, local_dir_header_ofs, local_dir_header, + sizeof(local_dir_header)) != sizeof(local_dir_header)) + return MZ_FALSE; + + if (!mz_zip_writer_add_to_central_dir( + pZip, pArchive_name, (mz_uint16)archive_name_size, NULL, 0, pComment, + comment_size, uncomp_size, comp_size, uncomp_crc32, method, 0, + dos_time, dos_date, local_dir_header_ofs, ext_attributes)) + return MZ_FALSE; + + pZip->m_total_files++; + pZip->m_archive_size = cur_archive_file_ofs; + + return MZ_TRUE; +} +#endif // #ifndef MINIZ_NO_STDIO + +mz_bool mz_zip_writer_add_from_zip_reader(mz_zip_archive *pZip, + mz_zip_archive *pSource_zip, + mz_uint file_index) { + mz_uint n, bit_flags, num_alignment_padding_bytes; + mz_uint64 comp_bytes_remaining, local_dir_header_ofs; + mz_uint64 cur_src_file_ofs, cur_dst_file_ofs; + mz_uint32 + local_header_u32[(MZ_ZIP_LOCAL_DIR_HEADER_SIZE + sizeof(mz_uint32) - 1) / + sizeof(mz_uint32)]; + mz_uint8 *pLocal_header = (mz_uint8 *)local_header_u32; + mz_uint8 central_header[MZ_ZIP_CENTRAL_DIR_HEADER_SIZE]; + size_t orig_central_dir_size; + mz_zip_internal_state *pState; + void *pBuf; + const mz_uint8 *pSrc_central_header; + + if ((!pZip) || (!pZip->m_pState) || (pZip->m_zip_mode != MZ_ZIP_MODE_WRITING)) + return MZ_FALSE; + if (NULL == + (pSrc_central_header = mz_zip_reader_get_cdh(pSource_zip, file_index))) + return MZ_FALSE; + pState = pZip->m_pState; + + num_alignment_padding_bytes = + mz_zip_writer_compute_padding_needed_for_file_alignment(pZip); + + // no zip64 support yet + if ((pZip->m_total_files == 0xFFFF) || + ((pZip->m_archive_size + num_alignment_padding_bytes + + MZ_ZIP_LOCAL_DIR_HEADER_SIZE + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE) > + 0xFFFFFFFF)) + return MZ_FALSE; + + cur_src_file_ofs = + MZ_READ_LE32(pSrc_central_header + MZ_ZIP_CDH_LOCAL_HEADER_OFS); + cur_dst_file_ofs = pZip->m_archive_size; + + if (pSource_zip->m_pRead(pSource_zip->m_pIO_opaque, cur_src_file_ofs, + pLocal_header, MZ_ZIP_LOCAL_DIR_HEADER_SIZE) != + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) + return MZ_FALSE; + if (MZ_READ_LE32(pLocal_header) != MZ_ZIP_LOCAL_DIR_HEADER_SIG) + return MZ_FALSE; + cur_src_file_ofs += MZ_ZIP_LOCAL_DIR_HEADER_SIZE; + + if (!mz_zip_writer_write_zeros(pZip, cur_dst_file_ofs, + num_alignment_padding_bytes)) + return MZ_FALSE; + cur_dst_file_ofs += num_alignment_padding_bytes; + local_dir_header_ofs = cur_dst_file_ofs; + if (pZip->m_file_offset_alignment) { + MZ_ASSERT((local_dir_header_ofs & (pZip->m_file_offset_alignment - 1)) == + 0); + } + + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_dst_file_ofs, pLocal_header, + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) != + MZ_ZIP_LOCAL_DIR_HEADER_SIZE) + return MZ_FALSE; + cur_dst_file_ofs += MZ_ZIP_LOCAL_DIR_HEADER_SIZE; + + n = MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_FILENAME_LEN_OFS) + + MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_EXTRA_LEN_OFS); + comp_bytes_remaining = + n + MZ_READ_LE32(pSrc_central_header + MZ_ZIP_CDH_COMPRESSED_SIZE_OFS); + + if (NULL == (pBuf = pZip->m_pAlloc( + pZip->m_pAlloc_opaque, 1, + (size_t)MZ_MAX(sizeof(mz_uint32) * 4, + MZ_MIN((mz_uint)MZ_ZIP_MAX_IO_BUF_SIZE, + comp_bytes_remaining))))) + return MZ_FALSE; + + while (comp_bytes_remaining) { + n = (mz_uint)MZ_MIN((mz_uint)MZ_ZIP_MAX_IO_BUF_SIZE, comp_bytes_remaining); + if (pSource_zip->m_pRead(pSource_zip->m_pIO_opaque, cur_src_file_ofs, pBuf, + n) != n) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pBuf); + return MZ_FALSE; + } + cur_src_file_ofs += n; + + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_dst_file_ofs, pBuf, n) != n) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pBuf); + return MZ_FALSE; + } + cur_dst_file_ofs += n; + + comp_bytes_remaining -= n; + } + + bit_flags = MZ_READ_LE16(pLocal_header + MZ_ZIP_LDH_BIT_FLAG_OFS); + if (bit_flags & 8) { + // Copy data descriptor + if (pSource_zip->m_pRead(pSource_zip->m_pIO_opaque, cur_src_file_ofs, pBuf, + sizeof(mz_uint32) * 4) != sizeof(mz_uint32) * 4) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pBuf); + return MZ_FALSE; + } + + n = sizeof(mz_uint32) * ((MZ_READ_LE32(pBuf) == 0x08074b50) ? 4 : 3); + if (pZip->m_pWrite(pZip->m_pIO_opaque, cur_dst_file_ofs, pBuf, n) != n) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pBuf); + return MZ_FALSE; + } + + cur_src_file_ofs += n; + cur_dst_file_ofs += n; + } + pZip->m_pFree(pZip->m_pAlloc_opaque, pBuf); + + // no zip64 support yet + if (cur_dst_file_ofs > 0xFFFFFFFF) return MZ_FALSE; + + orig_central_dir_size = pState->m_central_dir.m_size; + + memcpy(central_header, pSrc_central_header, MZ_ZIP_CENTRAL_DIR_HEADER_SIZE); + MZ_WRITE_LE32(central_header + MZ_ZIP_CDH_LOCAL_HEADER_OFS, + local_dir_header_ofs); + if (!mz_zip_array_push_back(pZip, &pState->m_central_dir, central_header, + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE)) + return MZ_FALSE; + + n = MZ_READ_LE16(pSrc_central_header + MZ_ZIP_CDH_FILENAME_LEN_OFS) + + MZ_READ_LE16(pSrc_central_header + MZ_ZIP_CDH_EXTRA_LEN_OFS) + + MZ_READ_LE16(pSrc_central_header + MZ_ZIP_CDH_COMMENT_LEN_OFS); + if (!mz_zip_array_push_back( + pZip, &pState->m_central_dir, + pSrc_central_header + MZ_ZIP_CENTRAL_DIR_HEADER_SIZE, n)) { + mz_zip_array_resize(pZip, &pState->m_central_dir, orig_central_dir_size, + MZ_FALSE); + return MZ_FALSE; + } + + if (pState->m_central_dir.m_size > 0xFFFFFFFF) return MZ_FALSE; + n = (mz_uint32)orig_central_dir_size; + if (!mz_zip_array_push_back(pZip, &pState->m_central_dir_offsets, &n, 1)) { + mz_zip_array_resize(pZip, &pState->m_central_dir, orig_central_dir_size, + MZ_FALSE); + return MZ_FALSE; + } + + pZip->m_total_files++; + pZip->m_archive_size = cur_dst_file_ofs; + + return MZ_TRUE; +} + +mz_bool mz_zip_writer_finalize_archive(mz_zip_archive *pZip) { + mz_zip_internal_state *pState; + mz_uint64 central_dir_ofs, central_dir_size; + mz_uint8 hdr[MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE]; + + if ((!pZip) || (!pZip->m_pState) || (pZip->m_zip_mode != MZ_ZIP_MODE_WRITING)) + return MZ_FALSE; + + pState = pZip->m_pState; + + // no zip64 support yet + if ((pZip->m_total_files > 0xFFFF) || + ((pZip->m_archive_size + pState->m_central_dir.m_size + + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIZE) > 0xFFFFFFFF)) + return MZ_FALSE; + + central_dir_ofs = 0; + central_dir_size = 0; + if (pZip->m_total_files) { + // Write central directory + central_dir_ofs = pZip->m_archive_size; + central_dir_size = pState->m_central_dir.m_size; + pZip->m_central_directory_file_ofs = central_dir_ofs; + if (pZip->m_pWrite(pZip->m_pIO_opaque, central_dir_ofs, + pState->m_central_dir.m_p, + (size_t)central_dir_size) != central_dir_size) + return MZ_FALSE; + pZip->m_archive_size += central_dir_size; + } + + // Write end of central directory record + MZ_CLEAR_OBJ(hdr); + MZ_WRITE_LE32(hdr + MZ_ZIP_ECDH_SIG_OFS, + MZ_ZIP_END_OF_CENTRAL_DIR_HEADER_SIG); + MZ_WRITE_LE16(hdr + MZ_ZIP_ECDH_CDIR_NUM_ENTRIES_ON_DISK_OFS, + pZip->m_total_files); + MZ_WRITE_LE16(hdr + MZ_ZIP_ECDH_CDIR_TOTAL_ENTRIES_OFS, pZip->m_total_files); + MZ_WRITE_LE32(hdr + MZ_ZIP_ECDH_CDIR_SIZE_OFS, central_dir_size); + MZ_WRITE_LE32(hdr + MZ_ZIP_ECDH_CDIR_OFS_OFS, central_dir_ofs); + + if (pZip->m_pWrite(pZip->m_pIO_opaque, pZip->m_archive_size, hdr, + sizeof(hdr)) != sizeof(hdr)) + return MZ_FALSE; +#ifndef MINIZ_NO_STDIO + if ((pState->m_pFile) && (MZ_FFLUSH(pState->m_pFile) == EOF)) return MZ_FALSE; +#endif // #ifndef MINIZ_NO_STDIO + + pZip->m_archive_size += sizeof(hdr); + + pZip->m_zip_mode = MZ_ZIP_MODE_WRITING_HAS_BEEN_FINALIZED; + return MZ_TRUE; +} + +mz_bool mz_zip_writer_finalize_heap_archive(mz_zip_archive *pZip, void **pBuf, + size_t *pSize) { + if ((!pZip) || (!pZip->m_pState) || (!pBuf) || (!pSize)) return MZ_FALSE; + if (pZip->m_pWrite != mz_zip_heap_write_func) return MZ_FALSE; + if (!mz_zip_writer_finalize_archive(pZip)) return MZ_FALSE; + + *pBuf = pZip->m_pState->m_pMem; + *pSize = pZip->m_pState->m_mem_size; + pZip->m_pState->m_pMem = NULL; + pZip->m_pState->m_mem_size = pZip->m_pState->m_mem_capacity = 0; + return MZ_TRUE; +} + +mz_bool mz_zip_writer_end(mz_zip_archive *pZip) { + mz_zip_internal_state *pState; + mz_bool status = MZ_TRUE; + if ((!pZip) || (!pZip->m_pState) || (!pZip->m_pAlloc) || (!pZip->m_pFree) || + ((pZip->m_zip_mode != MZ_ZIP_MODE_WRITING) && + (pZip->m_zip_mode != MZ_ZIP_MODE_WRITING_HAS_BEEN_FINALIZED))) + return MZ_FALSE; + + pState = pZip->m_pState; + pZip->m_pState = NULL; + mz_zip_array_clear(pZip, &pState->m_central_dir); + mz_zip_array_clear(pZip, &pState->m_central_dir_offsets); + mz_zip_array_clear(pZip, &pState->m_sorted_central_dir_offsets); + +#ifndef MINIZ_NO_STDIO + if (pState->m_pFile) { + MZ_FCLOSE(pState->m_pFile); + pState->m_pFile = NULL; + } +#endif // #ifndef MINIZ_NO_STDIO + + if ((pZip->m_pWrite == mz_zip_heap_write_func) && (pState->m_pMem)) { + pZip->m_pFree(pZip->m_pAlloc_opaque, pState->m_pMem); + pState->m_pMem = NULL; + } + + pZip->m_pFree(pZip->m_pAlloc_opaque, pState); + pZip->m_zip_mode = MZ_ZIP_MODE_INVALID; + return status; +} + +#ifndef MINIZ_NO_STDIO +mz_bool mz_zip_add_mem_to_archive_file_in_place( + const char *pZip_filename, const char *pArchive_name, const void *pBuf, + size_t buf_size, const void *pComment, mz_uint16 comment_size, + mz_uint level_and_flags) { + mz_bool status, created_new_archive = MZ_FALSE; + mz_zip_archive zip_archive; + struct MZ_FILE_STAT_STRUCT file_stat; + MZ_CLEAR_OBJ(zip_archive); + if ((int)level_and_flags < 0) level_and_flags = MZ_DEFAULT_LEVEL; + if ((!pZip_filename) || (!pArchive_name) || ((buf_size) && (!pBuf)) || + ((comment_size) && (!pComment)) || + ((level_and_flags & 0xF) > MZ_UBER_COMPRESSION)) + return MZ_FALSE; + if (!mz_zip_writer_validate_archive_name(pArchive_name)) return MZ_FALSE; + if (MZ_FILE_STAT(pZip_filename, &file_stat) != 0) { + // Create a new archive. + if (!mz_zip_writer_init_file(&zip_archive, pZip_filename, 0)) + return MZ_FALSE; + created_new_archive = MZ_TRUE; + } else { + // Append to an existing archive. + if (!mz_zip_reader_init_file( + &zip_archive, pZip_filename, + level_and_flags | MZ_ZIP_FLAG_DO_NOT_SORT_CENTRAL_DIRECTORY)) + return MZ_FALSE; + if (!mz_zip_writer_init_from_reader(&zip_archive, pZip_filename)) { + mz_zip_reader_end(&zip_archive); + return MZ_FALSE; + } + } + status = + mz_zip_writer_add_mem_ex(&zip_archive, pArchive_name, pBuf, buf_size, + pComment, comment_size, level_and_flags, 0, 0); + // Always finalize, even if adding failed for some reason, so we have a valid + // central directory. (This may not always succeed, but we can try.) + if (!mz_zip_writer_finalize_archive(&zip_archive)) status = MZ_FALSE; + if (!mz_zip_writer_end(&zip_archive)) status = MZ_FALSE; + if ((!status) && (created_new_archive)) { + // It's a new archive and something went wrong, so just delete it. + int ignoredStatus = MZ_DELETE_FILE(pZip_filename); + (void)ignoredStatus; + } + return status; +} + +void *mz_zip_extract_archive_file_to_heap(const char *pZip_filename, + const char *pArchive_name, + size_t *pSize, mz_uint flags) { + int file_index; + mz_zip_archive zip_archive; + void *p = NULL; + + if (pSize) *pSize = 0; + + if ((!pZip_filename) || (!pArchive_name)) return NULL; + + MZ_CLEAR_OBJ(zip_archive); + if (!mz_zip_reader_init_file( + &zip_archive, pZip_filename, + flags | MZ_ZIP_FLAG_DO_NOT_SORT_CENTRAL_DIRECTORY)) + return NULL; + + if ((file_index = mz_zip_reader_locate_file(&zip_archive, pArchive_name, NULL, + flags)) >= 0) + p = mz_zip_reader_extract_to_heap(&zip_archive, file_index, pSize, flags); + + mz_zip_reader_end(&zip_archive); + return p; +} + +#endif // #ifndef MINIZ_NO_STDIO + +#endif // #ifndef MINIZ_NO_ARCHIVE_WRITING_APIS + +#endif // #ifndef MINIZ_NO_ARCHIVE_APIS + +#ifdef __cplusplus +} +#endif + +#ifdef _MSC_VER +#pragma warning(pop) +#endif + +#endif // MINIZ_HEADER_FILE_ONLY + +/* + This is free and unencumbered software released into the public domain. + + Anyone is free to copy, modify, publish, use, compile, sell, or + distribute this software, either in source code form or as a compiled + binary, for any purpose, commercial or non-commercial, and by any + means. + + In jurisdictions that recognize copyright laws, the author or authors + of this software dedicate any and all copyright interest in the + software to the public domain. We make this dedication for the benefit + of the public at large and to the detriment of our heirs and + successors. We intend this dedication to be an overt act of + relinquishment in perpetuity of all present and future rights to this + software under copyright law. + + THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. + IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR + OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + OTHER DEALINGS IN THE SOFTWARE. + + For more information, please refer to +*/ + +// ---------------------- end of miniz ---------------------------------------- + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +} // namespace miniz +#else + +// Reuse MINIZ_LITTE_ENDIAN macro + +#if defined(__sparcv9) +// Big endian +#else +#if (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) || MINIZ_X86_OR_X64_CPU +// Set MINIZ_LITTLE_ENDIAN to 1 if the processor is little endian. +#define MINIZ_LITTLE_ENDIAN 1 +#endif +#endif + +#endif // TINYEXR_USE_MINIZ + +// static bool IsBigEndian(void) { +// union { +// unsigned int i; +// char c[4]; +// } bint = {0x01020304}; +// +// return bint.c[0] == 1; +//} + +static void SetErrorMessage(const std::string &msg, const char **err) { + if (err) { +#ifdef _WIN32 + (*err) = _strdup(msg.c_str()); +#else + (*err) = strdup(msg.c_str()); +#endif + } +} + +static const int kEXRVersionSize = 8; + +static void cpy2(unsigned short *dst_val, const unsigned short *src_val) { + unsigned char *dst = reinterpret_cast(dst_val); + const unsigned char *src = reinterpret_cast(src_val); + + dst[0] = src[0]; + dst[1] = src[1]; +} + +static void swap2(unsigned short *val) { +#ifdef MINIZ_LITTLE_ENDIAN + (void)val; +#else + unsigned short tmp = *val; + unsigned char *dst = reinterpret_cast(val); + unsigned char *src = reinterpret_cast(&tmp); + + dst[0] = src[1]; + dst[1] = src[0]; +#endif +} + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wunused-function" +#endif + +#ifdef __GNUC__ +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wunused-function" +#endif +static void cpy4(int *dst_val, const int *src_val) { + unsigned char *dst = reinterpret_cast(dst_val); + const unsigned char *src = reinterpret_cast(src_val); + + dst[0] = src[0]; + dst[1] = src[1]; + dst[2] = src[2]; + dst[3] = src[3]; +} + +static void cpy4(unsigned int *dst_val, const unsigned int *src_val) { + unsigned char *dst = reinterpret_cast(dst_val); + const unsigned char *src = reinterpret_cast(src_val); + + dst[0] = src[0]; + dst[1] = src[1]; + dst[2] = src[2]; + dst[3] = src[3]; +} + +static void cpy4(float *dst_val, const float *src_val) { + unsigned char *dst = reinterpret_cast(dst_val); + const unsigned char *src = reinterpret_cast(src_val); + + dst[0] = src[0]; + dst[1] = src[1]; + dst[2] = src[2]; + dst[3] = src[3]; +} +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +#ifdef __GNUC__ +#pragma GCC diagnostic pop +#endif + +static void swap4(unsigned int *val) { +#ifdef MINIZ_LITTLE_ENDIAN + (void)val; +#else + unsigned int tmp = *val; + unsigned char *dst = reinterpret_cast(val); + unsigned char *src = reinterpret_cast(&tmp); + + dst[0] = src[3]; + dst[1] = src[2]; + dst[2] = src[1]; + dst[3] = src[0]; +#endif +} + +#if 0 +static void cpy8(tinyexr::tinyexr_uint64 *dst_val, const tinyexr::tinyexr_uint64 *src_val) { + unsigned char *dst = reinterpret_cast(dst_val); + const unsigned char *src = reinterpret_cast(src_val); + + dst[0] = src[0]; + dst[1] = src[1]; + dst[2] = src[2]; + dst[3] = src[3]; + dst[4] = src[4]; + dst[5] = src[5]; + dst[6] = src[6]; + dst[7] = src[7]; +} +#endif + +static void swap8(tinyexr::tinyexr_uint64 *val) { +#ifdef MINIZ_LITTLE_ENDIAN + (void)val; +#else + tinyexr::tinyexr_uint64 tmp = (*val); + unsigned char *dst = reinterpret_cast(val); + unsigned char *src = reinterpret_cast(&tmp); + + dst[0] = src[7]; + dst[1] = src[6]; + dst[2] = src[5]; + dst[3] = src[4]; + dst[4] = src[3]; + dst[5] = src[2]; + dst[6] = src[1]; + dst[7] = src[0]; +#endif +} + +// https://gist.github.com/rygorous/2156668 +// Reuse MINIZ_LITTLE_ENDIAN flag from miniz. +union FP32 { + unsigned int u; + float f; + struct { +#if MINIZ_LITTLE_ENDIAN + unsigned int Mantissa : 23; + unsigned int Exponent : 8; + unsigned int Sign : 1; +#else + unsigned int Sign : 1; + unsigned int Exponent : 8; + unsigned int Mantissa : 23; +#endif + } s; +}; + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wpadded" +#endif + +union FP16 { + unsigned short u; + struct { +#if MINIZ_LITTLE_ENDIAN + unsigned int Mantissa : 10; + unsigned int Exponent : 5; + unsigned int Sign : 1; +#else + unsigned int Sign : 1; + unsigned int Exponent : 5; + unsigned int Mantissa : 10; +#endif + } s; +}; + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +static FP32 half_to_float(FP16 h) { + static const FP32 magic = {113 << 23}; + static const unsigned int shifted_exp = 0x7c00 + << 13; // exponent mask after shift + FP32 o; + + o.u = (h.u & 0x7fffU) << 13U; // exponent/mantissa bits + unsigned int exp_ = shifted_exp & o.u; // just the exponent + o.u += (127 - 15) << 23; // exponent adjust + + // handle exponent special cases + if (exp_ == shifted_exp) // Inf/NaN? + o.u += (128 - 16) << 23; // extra exp adjust + else if (exp_ == 0) // Zero/Denormal? + { + o.u += 1 << 23; // extra exp adjust + o.f -= magic.f; // renormalize + } + + o.u |= (h.u & 0x8000U) << 16U; // sign bit + return o; +} + +static FP16 float_to_half_full(FP32 f) { + FP16 o = {0}; + + // Based on ISPC reference code (with minor modifications) + if (f.s.Exponent == 0) // Signed zero/denormal (which will underflow) + o.s.Exponent = 0; + else if (f.s.Exponent == 255) // Inf or NaN (all exponent bits set) + { + o.s.Exponent = 31; + o.s.Mantissa = f.s.Mantissa ? 0x200 : 0; // NaN->qNaN and Inf->Inf + } else // Normalized number + { + // Exponent unbias the single, then bias the halfp + int newexp = f.s.Exponent - 127 + 15; + if (newexp >= 31) // Overflow, return signed infinity + o.s.Exponent = 31; + else if (newexp <= 0) // Underflow + { + if ((14 - newexp) <= 24) // Mantissa might be non-zero + { + unsigned int mant = f.s.Mantissa | 0x800000; // Hidden 1 bit + o.s.Mantissa = mant >> (14 - newexp); + if ((mant >> (13 - newexp)) & 1) // Check for rounding + o.u++; // Round, might overflow into exp bit, but this is OK + } + } else { + o.s.Exponent = static_cast(newexp); + o.s.Mantissa = f.s.Mantissa >> 13; + if (f.s.Mantissa & 0x1000) // Check for rounding + o.u++; // Round, might overflow to inf, this is OK + } + } + + o.s.Sign = f.s.Sign; + return o; +} + +// NOTE: From OpenEXR code +// #define IMF_INCREASING_Y 0 +// #define IMF_DECREASING_Y 1 +// #define IMF_RAMDOM_Y 2 +// +// #define IMF_NO_COMPRESSION 0 +// #define IMF_RLE_COMPRESSION 1 +// #define IMF_ZIPS_COMPRESSION 2 +// #define IMF_ZIP_COMPRESSION 3 +// #define IMF_PIZ_COMPRESSION 4 +// #define IMF_PXR24_COMPRESSION 5 +// #define IMF_B44_COMPRESSION 6 +// #define IMF_B44A_COMPRESSION 7 + +#ifdef __clang__ +#pragma clang diagnostic push + +#if __has_warning("-Wzero-as-null-pointer-constant") +#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant" +#endif + +#endif + +static const char *ReadString(std::string *s, const char *ptr, size_t len) { + // Read untile NULL(\0). + const char *p = ptr; + const char *q = ptr; + while ((size_t(q - ptr) < len) && (*q) != 0) { + q++; + } + + if (size_t(q - ptr) >= len) { + (*s) = std::string(); + return NULL; + } + + (*s) = std::string(p, q); + + return q + 1; // skip '\0' +} + +static bool ReadAttribute(std::string *name, std::string *type, + std::vector *data, size_t *marker_size, + const char *marker, size_t size) { + size_t name_len = strnlen(marker, size); + if (name_len == size) { + // String does not have a terminating character. + return false; + } + *name = std::string(marker, name_len); + + marker += name_len + 1; + size -= name_len + 1; + + size_t type_len = strnlen(marker, size); + if (type_len == size) { + return false; + } + *type = std::string(marker, type_len); + + marker += type_len + 1; + size -= type_len + 1; + + if (size < sizeof(uint32_t)) { + return false; + } + + uint32_t data_len; + memcpy(&data_len, marker, sizeof(uint32_t)); + tinyexr::swap4(reinterpret_cast(&data_len)); + + if (data_len == 0) { + if ((*type).compare("string") == 0) { + // Accept empty string attribute. + + marker += sizeof(uint32_t); + size -= sizeof(uint32_t); + + *marker_size = name_len + 1 + type_len + 1 + sizeof(uint32_t); + + data->resize(1); + (*data)[0] = '\0'; + + return true; + } else { + return false; + } + } + + marker += sizeof(uint32_t); + size -= sizeof(uint32_t); + + if (size < data_len) { + return false; + } + + data->resize(static_cast(data_len)); + memcpy(&data->at(0), marker, static_cast(data_len)); + + *marker_size = name_len + 1 + type_len + 1 + sizeof(uint32_t) + data_len; + return true; +} + +static void WriteAttributeToMemory(std::vector *out, + const char *name, const char *type, + const unsigned char *data, int len) { + out->insert(out->end(), name, name + strlen(name) + 1); + out->insert(out->end(), type, type + strlen(type) + 1); + + int outLen = len; + tinyexr::swap4(reinterpret_cast(&outLen)); + out->insert(out->end(), reinterpret_cast(&outLen), + reinterpret_cast(&outLen) + sizeof(int)); + out->insert(out->end(), data, data + len); +} + +struct ChannelInfo { + std::string name; // less than 255 bytes long + int pixel_type; + int x_sampling; + int y_sampling; + unsigned char p_linear; + unsigned char pad[3]; +}; + +struct HeaderInfo { + std::vector channels; + std::vector attributes; + + int data_window[4]; + int line_order; + int display_window[4]; + float screen_window_center[2]; + float screen_window_width; + float pixel_aspect_ratio; + + int chunk_count; + + // Tiled format + int tile_size_x; + int tile_size_y; + int tile_level_mode; + int tile_rounding_mode; + + unsigned int header_len; + + int compression_type; + + void clear() { + channels.clear(); + attributes.clear(); + + data_window[0] = 0; + data_window[1] = 0; + data_window[2] = 0; + data_window[3] = 0; + line_order = 0; + display_window[0] = 0; + display_window[1] = 0; + display_window[2] = 0; + display_window[3] = 0; + screen_window_center[0] = 0.0f; + screen_window_center[1] = 0.0f; + screen_window_width = 0.0f; + pixel_aspect_ratio = 0.0f; + + chunk_count = 0; + + // Tiled format + tile_size_x = 0; + tile_size_y = 0; + tile_level_mode = 0; + tile_rounding_mode = 0; + + header_len = 0; + compression_type = 0; + } +}; + +static bool ReadChannelInfo(std::vector &channels, + const std::vector &data) { + const char *p = reinterpret_cast(&data.at(0)); + + for (;;) { + if ((*p) == 0) { + break; + } + ChannelInfo info; + + tinyexr_int64 data_len = static_cast(data.size()) - + (p - reinterpret_cast(data.data())); + if (data_len < 0) { + return false; + } + + p = ReadString(&info.name, p, size_t(data_len)); + if ((p == NULL) && (info.name.empty())) { + // Buffer overrun. Issue #51. + return false; + } + + const unsigned char *data_end = + reinterpret_cast(p) + 16; + if (data_end >= (data.data() + data.size())) { + return false; + } + + memcpy(&info.pixel_type, p, sizeof(int)); + p += 4; + info.p_linear = static_cast(p[0]); // uchar + p += 1 + 3; // reserved: uchar[3] + memcpy(&info.x_sampling, p, sizeof(int)); // int + p += 4; + memcpy(&info.y_sampling, p, sizeof(int)); // int + p += 4; + + tinyexr::swap4(reinterpret_cast(&info.pixel_type)); + tinyexr::swap4(reinterpret_cast(&info.x_sampling)); + tinyexr::swap4(reinterpret_cast(&info.y_sampling)); + + channels.push_back(info); + } + + return true; +} + +static void WriteChannelInfo(std::vector &data, + const std::vector &channels) { + size_t sz = 0; + + // Calculate total size. + for (size_t c = 0; c < channels.size(); c++) { + sz += strlen(channels[c].name.c_str()) + 1; // +1 for \0 + sz += 16; // 4 * int + } + data.resize(sz + 1); + + unsigned char *p = &data.at(0); + + for (size_t c = 0; c < channels.size(); c++) { + memcpy(p, channels[c].name.c_str(), strlen(channels[c].name.c_str())); + p += strlen(channels[c].name.c_str()); + (*p) = '\0'; + p++; + + int pixel_type = channels[c].pixel_type; + int x_sampling = channels[c].x_sampling; + int y_sampling = channels[c].y_sampling; + tinyexr::swap4(reinterpret_cast(&pixel_type)); + tinyexr::swap4(reinterpret_cast(&x_sampling)); + tinyexr::swap4(reinterpret_cast(&y_sampling)); + + memcpy(p, &pixel_type, sizeof(int)); + p += sizeof(int); + + (*p) = channels[c].p_linear; + p += 4; + + memcpy(p, &x_sampling, sizeof(int)); + p += sizeof(int); + + memcpy(p, &y_sampling, sizeof(int)); + p += sizeof(int); + } + + (*p) = '\0'; +} + +static void CompressZip(unsigned char *dst, + tinyexr::tinyexr_uint64 &compressedSize, + const unsigned char *src, unsigned long src_size) { + std::vector tmpBuf(src_size); + + // + // Apply EXR-specific? postprocess. Grabbed from OpenEXR's + // ImfZipCompressor.cpp + // + + // + // Reorder the pixel data. + // + + const char *srcPtr = reinterpret_cast(src); + + { + char *t1 = reinterpret_cast(&tmpBuf.at(0)); + char *t2 = reinterpret_cast(&tmpBuf.at(0)) + (src_size + 1) / 2; + const char *stop = srcPtr + src_size; + + for (;;) { + if (srcPtr < stop) + *(t1++) = *(srcPtr++); + else + break; + + if (srcPtr < stop) + *(t2++) = *(srcPtr++); + else + break; + } + } + + // + // Predictor. + // + + { + unsigned char *t = &tmpBuf.at(0) + 1; + unsigned char *stop = &tmpBuf.at(0) + src_size; + int p = t[-1]; + + while (t < stop) { + int d = int(t[0]) - p + (128 + 256); + p = t[0]; + t[0] = static_cast(d); + ++t; + } + } + +#if TINYEXR_USE_MINIZ + // + // Compress the data using miniz + // + + miniz::mz_ulong outSize = miniz::mz_compressBound(src_size); + int ret = miniz::mz_compress( + dst, &outSize, static_cast(&tmpBuf.at(0)), + src_size); + TEXR_ASSERT(ret == miniz::MZ_OK); + (void)ret; + + compressedSize = outSize; +#else + uLong outSize = compressBound(static_cast(src_size)); + int ret = compress(dst, &outSize, static_cast(&tmpBuf.at(0)), + src_size); + TEXR_ASSERT(ret == Z_OK); + + compressedSize = outSize; +#endif + + // Use uncompressed data when compressed data is larger than uncompressed. + // (Issue 40) + if (compressedSize >= src_size) { + compressedSize = src_size; + memcpy(dst, src, src_size); + } +} + +static bool DecompressZip(unsigned char *dst, + unsigned long *uncompressed_size /* inout */, + const unsigned char *src, unsigned long src_size) { + if ((*uncompressed_size) == src_size) { + // Data is not compressed(Issue 40). + memcpy(dst, src, src_size); + return true; + } + std::vector tmpBuf(*uncompressed_size); + +#if TINYEXR_USE_MINIZ + int ret = + miniz::mz_uncompress(&tmpBuf.at(0), uncompressed_size, src, src_size); + if (miniz::MZ_OK != ret) { + return false; + } +#else + int ret = uncompress(&tmpBuf.at(0), uncompressed_size, src, src_size); + if (Z_OK != ret) { + return false; + } +#endif + + // + // Apply EXR-specific? postprocess. Grabbed from OpenEXR's + // ImfZipCompressor.cpp + // + + // Predictor. + { + unsigned char *t = &tmpBuf.at(0) + 1; + unsigned char *stop = &tmpBuf.at(0) + (*uncompressed_size); + + while (t < stop) { + int d = int(t[-1]) + int(t[0]) - 128; + t[0] = static_cast(d); + ++t; + } + } + + // Reorder the pixel data. + { + const char *t1 = reinterpret_cast(&tmpBuf.at(0)); + const char *t2 = reinterpret_cast(&tmpBuf.at(0)) + + (*uncompressed_size + 1) / 2; + char *s = reinterpret_cast(dst); + char *stop = s + (*uncompressed_size); + + for (;;) { + if (s < stop) + *(s++) = *(t1++); + else + break; + + if (s < stop) + *(s++) = *(t2++); + else + break; + } + } + + return true; +} + +// RLE code from OpenEXR -------------------------------------- + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wsign-conversion" +#if __has_warning("-Wextra-semi-stmt") +#pragma clang diagnostic ignored "-Wextra-semi-stmt" +#endif +#endif + +#ifdef _MSC_VER +#pragma warning(push) +#pragma warning(disable : 4204) // nonstandard extension used : non-constant + // aggregate initializer (also supported by GNU + // C and C99, so no big deal) +#pragma warning(disable : 4244) // 'initializing': conversion from '__int64' to + // 'int', possible loss of data +#pragma warning(disable : 4267) // 'argument': conversion from '__int64' to + // 'int', possible loss of data +#pragma warning(disable : 4996) // 'strdup': The POSIX name for this item is + // deprecated. Instead, use the ISO C and C++ + // conformant name: _strdup. +#endif + +const int MIN_RUN_LENGTH = 3; +const int MAX_RUN_LENGTH = 127; + +// +// Compress an array of bytes, using run-length encoding, +// and return the length of the compressed data. +// + +static int rleCompress(int inLength, const char in[], signed char out[]) { + const char *inEnd = in + inLength; + const char *runStart = in; + const char *runEnd = in + 1; + signed char *outWrite = out; + + while (runStart < inEnd) { + while (runEnd < inEnd && *runStart == *runEnd && + runEnd - runStart - 1 < MAX_RUN_LENGTH) { + ++runEnd; + } + + if (runEnd - runStart >= MIN_RUN_LENGTH) { + // + // Compressable run + // + + *outWrite++ = static_cast(runEnd - runStart) - 1; + *outWrite++ = *(reinterpret_cast(runStart)); + runStart = runEnd; + } else { + // + // Uncompressable run + // + + while (runEnd < inEnd && + ((runEnd + 1 >= inEnd || *runEnd != *(runEnd + 1)) || + (runEnd + 2 >= inEnd || *(runEnd + 1) != *(runEnd + 2))) && + runEnd - runStart < MAX_RUN_LENGTH) { + ++runEnd; + } + + *outWrite++ = static_cast(runStart - runEnd); + + while (runStart < runEnd) { + *outWrite++ = *(reinterpret_cast(runStart++)); + } + } + + ++runEnd; + } + + return static_cast(outWrite - out); +} + +// +// Uncompress an array of bytes compressed with rleCompress(). +// Returns the length of the oncompressed data, or 0 if the +// length of the uncompressed data would be more than maxLength. +// + +static int rleUncompress(int inLength, int maxLength, const signed char in[], + char out[]) { + char *outStart = out; + + while (inLength > 0) { + if (*in < 0) { + int count = -(static_cast(*in++)); + inLength -= count + 1; + + // Fixes #116: Add bounds check to in buffer. + if ((0 > (maxLength -= count)) || (inLength < 0)) return 0; + + memcpy(out, in, count); + out += count; + in += count; + } else { + int count = *in++; + inLength -= 2; + + if (0 > (maxLength -= count + 1)) return 0; + + memset(out, *reinterpret_cast(in), count + 1); + out += count + 1; + + in++; + } + } + + return static_cast(out - outStart); +} + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +// End of RLE code from OpenEXR ----------------------------------- + +static void CompressRle(unsigned char *dst, + tinyexr::tinyexr_uint64 &compressedSize, + const unsigned char *src, unsigned long src_size) { + std::vector tmpBuf(src_size); + + // + // Apply EXR-specific? postprocess. Grabbed from OpenEXR's + // ImfRleCompressor.cpp + // + + // + // Reorder the pixel data. + // + + const char *srcPtr = reinterpret_cast(src); + + { + char *t1 = reinterpret_cast(&tmpBuf.at(0)); + char *t2 = reinterpret_cast(&tmpBuf.at(0)) + (src_size + 1) / 2; + const char *stop = srcPtr + src_size; + + for (;;) { + if (srcPtr < stop) + *(t1++) = *(srcPtr++); + else + break; + + if (srcPtr < stop) + *(t2++) = *(srcPtr++); + else + break; + } + } + + // + // Predictor. + // + + { + unsigned char *t = &tmpBuf.at(0) + 1; + unsigned char *stop = &tmpBuf.at(0) + src_size; + int p = t[-1]; + + while (t < stop) { + int d = int(t[0]) - p + (128 + 256); + p = t[0]; + t[0] = static_cast(d); + ++t; + } + } + + // outSize will be (srcSiz * 3) / 2 at max. + int outSize = rleCompress(static_cast(src_size), + reinterpret_cast(&tmpBuf.at(0)), + reinterpret_cast(dst)); + TEXR_ASSERT(outSize > 0); + + compressedSize = static_cast(outSize); + + // Use uncompressed data when compressed data is larger than uncompressed. + // (Issue 40) + if (compressedSize >= src_size) { + compressedSize = src_size; + memcpy(dst, src, src_size); + } +} + +static bool DecompressRle(unsigned char *dst, + const unsigned long uncompressed_size, + const unsigned char *src, unsigned long src_size) { + if (uncompressed_size == src_size) { + // Data is not compressed(Issue 40). + memcpy(dst, src, src_size); + return true; + } + + // Workaround for issue #112. + // TODO(syoyo): Add more robust out-of-bounds check in `rleUncompress`. + if (src_size <= 2) { + return false; + } + + std::vector tmpBuf(uncompressed_size); + + int ret = rleUncompress(static_cast(src_size), + static_cast(uncompressed_size), + reinterpret_cast(src), + reinterpret_cast(&tmpBuf.at(0))); + if (ret != static_cast(uncompressed_size)) { + return false; + } + + // + // Apply EXR-specific? postprocess. Grabbed from OpenEXR's + // ImfRleCompressor.cpp + // + + // Predictor. + { + unsigned char *t = &tmpBuf.at(0) + 1; + unsigned char *stop = &tmpBuf.at(0) + uncompressed_size; + + while (t < stop) { + int d = int(t[-1]) + int(t[0]) - 128; + t[0] = static_cast(d); + ++t; + } + } + + // Reorder the pixel data. + { + const char *t1 = reinterpret_cast(&tmpBuf.at(0)); + const char *t2 = reinterpret_cast(&tmpBuf.at(0)) + + (uncompressed_size + 1) / 2; + char *s = reinterpret_cast(dst); + char *stop = s + uncompressed_size; + + for (;;) { + if (s < stop) + *(s++) = *(t1++); + else + break; + + if (s < stop) + *(s++) = *(t2++); + else + break; + } + } + + return true; +} + +#if TINYEXR_USE_PIZ + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wc++11-long-long" +#pragma clang diagnostic ignored "-Wold-style-cast" +#pragma clang diagnostic ignored "-Wpadded" +#pragma clang diagnostic ignored "-Wsign-conversion" +#pragma clang diagnostic ignored "-Wc++11-extensions" +#pragma clang diagnostic ignored "-Wconversion" +#pragma clang diagnostic ignored "-Wc++98-compat-pedantic" + +#if __has_warning("-Wcast-qual") +#pragma clang diagnostic ignored "-Wcast-qual" +#endif + +#if __has_warning("-Wextra-semi-stmt") +#pragma clang diagnostic ignored "-Wextra-semi-stmt" +#endif + +#endif + +// +// PIZ compress/uncompress, based on OpenEXR's ImfPizCompressor.cpp +// +// ----------------------------------------------------------------- +// Copyright (c) 2004, Industrial Light & Magic, a division of Lucas +// Digital Ltd. LLC) +// (3 clause BSD license) +// + +struct PIZChannelData { + unsigned short *start; + unsigned short *end; + int nx; + int ny; + int ys; + int size; +}; + +//----------------------------------------------------------------------------- +// +// 16-bit Haar Wavelet encoding and decoding +// +// The source code in this file is derived from the encoding +// and decoding routines written by Christian Rouet for his +// PIZ image file format. +// +//----------------------------------------------------------------------------- + +// +// Wavelet basis functions without modulo arithmetic; they produce +// the best compression ratios when the wavelet-transformed data are +// Huffman-encoded, but the wavelet transform works only for 14-bit +// data (untransformed data values must be less than (1 << 14)). +// + +inline void wenc14(unsigned short a, unsigned short b, unsigned short &l, + unsigned short &h) { + short as = static_cast(a); + short bs = static_cast(b); + + short ms = (as + bs) >> 1; + short ds = as - bs; + + l = static_cast(ms); + h = static_cast(ds); +} + +inline void wdec14(unsigned short l, unsigned short h, unsigned short &a, + unsigned short &b) { + short ls = static_cast(l); + short hs = static_cast(h); + + int hi = hs; + int ai = ls + (hi & 1) + (hi >> 1); + + short as = static_cast(ai); + short bs = static_cast(ai - hi); + + a = static_cast(as); + b = static_cast(bs); +} + +// +// Wavelet basis functions with modulo arithmetic; they work with full +// 16-bit data, but Huffman-encoding the wavelet-transformed data doesn't +// compress the data quite as well. +// + +const int NBITS = 16; +const int A_OFFSET = 1 << (NBITS - 1); +const int M_OFFSET = 1 << (NBITS - 1); +const int MOD_MASK = (1 << NBITS) - 1; + +inline void wenc16(unsigned short a, unsigned short b, unsigned short &l, + unsigned short &h) { + int ao = (a + A_OFFSET) & MOD_MASK; + int m = ((ao + b) >> 1); + int d = ao - b; + + if (d < 0) m = (m + M_OFFSET) & MOD_MASK; + + d &= MOD_MASK; + + l = static_cast(m); + h = static_cast(d); +} + +inline void wdec16(unsigned short l, unsigned short h, unsigned short &a, + unsigned short &b) { + int m = l; + int d = h; + int bb = (m - (d >> 1)) & MOD_MASK; + int aa = (d + bb - A_OFFSET) & MOD_MASK; + b = static_cast(bb); + a = static_cast(aa); +} + +// +// 2D Wavelet encoding: +// + +static void wav2Encode( + unsigned short *in, // io: values are transformed in place + int nx, // i : x size + int ox, // i : x offset + int ny, // i : y size + int oy, // i : y offset + unsigned short mx) // i : maximum in[x][y] value +{ + bool w14 = (mx < (1 << 14)); + int n = (nx > ny) ? ny : nx; + int p = 1; // == 1 << level + int p2 = 2; // == 1 << (level+1) + + // + // Hierachical loop on smaller dimension n + // + + while (p2 <= n) { + unsigned short *py = in; + unsigned short *ey = in + oy * (ny - p2); + int oy1 = oy * p; + int oy2 = oy * p2; + int ox1 = ox * p; + int ox2 = ox * p2; + unsigned short i00, i01, i10, i11; + + // + // Y loop + // + + for (; py <= ey; py += oy2) { + unsigned short *px = py; + unsigned short *ex = py + ox * (nx - p2); + + // + // X loop + // + + for (; px <= ex; px += ox2) { + unsigned short *p01 = px + ox1; + unsigned short *p10 = px + oy1; + unsigned short *p11 = p10 + ox1; + + // + // 2D wavelet encoding + // + + if (w14) { + wenc14(*px, *p01, i00, i01); + wenc14(*p10, *p11, i10, i11); + wenc14(i00, i10, *px, *p10); + wenc14(i01, i11, *p01, *p11); + } else { + wenc16(*px, *p01, i00, i01); + wenc16(*p10, *p11, i10, i11); + wenc16(i00, i10, *px, *p10); + wenc16(i01, i11, *p01, *p11); + } + } + + // + // Encode (1D) odd column (still in Y loop) + // + + if (nx & p) { + unsigned short *p10 = px + oy1; + + if (w14) + wenc14(*px, *p10, i00, *p10); + else + wenc16(*px, *p10, i00, *p10); + + *px = i00; + } + } + + // + // Encode (1D) odd line (must loop in X) + // + + if (ny & p) { + unsigned short *px = py; + unsigned short *ex = py + ox * (nx - p2); + + for (; px <= ex; px += ox2) { + unsigned short *p01 = px + ox1; + + if (w14) + wenc14(*px, *p01, i00, *p01); + else + wenc16(*px, *p01, i00, *p01); + + *px = i00; + } + } + + // + // Next level + // + + p = p2; + p2 <<= 1; + } +} + +// +// 2D Wavelet decoding: +// + +static void wav2Decode( + unsigned short *in, // io: values are transformed in place + int nx, // i : x size + int ox, // i : x offset + int ny, // i : y size + int oy, // i : y offset + unsigned short mx) // i : maximum in[x][y] value +{ + bool w14 = (mx < (1 << 14)); + int n = (nx > ny) ? ny : nx; + int p = 1; + int p2; + + // + // Search max level + // + + while (p <= n) p <<= 1; + + p >>= 1; + p2 = p; + p >>= 1; + + // + // Hierarchical loop on smaller dimension n + // + + while (p >= 1) { + unsigned short *py = in; + unsigned short *ey = in + oy * (ny - p2); + int oy1 = oy * p; + int oy2 = oy * p2; + int ox1 = ox * p; + int ox2 = ox * p2; + unsigned short i00, i01, i10, i11; + + // + // Y loop + // + + for (; py <= ey; py += oy2) { + unsigned short *px = py; + unsigned short *ex = py + ox * (nx - p2); + + // + // X loop + // + + for (; px <= ex; px += ox2) { + unsigned short *p01 = px + ox1; + unsigned short *p10 = px + oy1; + unsigned short *p11 = p10 + ox1; + + // + // 2D wavelet decoding + // + + if (w14) { + wdec14(*px, *p10, i00, i10); + wdec14(*p01, *p11, i01, i11); + wdec14(i00, i01, *px, *p01); + wdec14(i10, i11, *p10, *p11); + } else { + wdec16(*px, *p10, i00, i10); + wdec16(*p01, *p11, i01, i11); + wdec16(i00, i01, *px, *p01); + wdec16(i10, i11, *p10, *p11); + } + } + + // + // Decode (1D) odd column (still in Y loop) + // + + if (nx & p) { + unsigned short *p10 = px + oy1; + + if (w14) + wdec14(*px, *p10, i00, *p10); + else + wdec16(*px, *p10, i00, *p10); + + *px = i00; + } + } + + // + // Decode (1D) odd line (must loop in X) + // + + if (ny & p) { + unsigned short *px = py; + unsigned short *ex = py + ox * (nx - p2); + + for (; px <= ex; px += ox2) { + unsigned short *p01 = px + ox1; + + if (w14) + wdec14(*px, *p01, i00, *p01); + else + wdec16(*px, *p01, i00, *p01); + + *px = i00; + } + } + + // + // Next level + // + + p2 = p; + p >>= 1; + } +} + +//----------------------------------------------------------------------------- +// +// 16-bit Huffman compression and decompression. +// +// The source code in this file is derived from the 8-bit +// Huffman compression and decompression routines written +// by Christian Rouet for his PIZ image file format. +// +//----------------------------------------------------------------------------- + +// Adds some modification for tinyexr. + +const int HUF_ENCBITS = 16; // literal (value) bit length +const int HUF_DECBITS = 14; // decoding bit size (>= 8) + +const int HUF_ENCSIZE = (1 << HUF_ENCBITS) + 1; // encoding table size +const int HUF_DECSIZE = 1 << HUF_DECBITS; // decoding table size +const int HUF_DECMASK = HUF_DECSIZE - 1; + +struct HufDec { // short code long code + //------------------------------- + int len : 8; // code length 0 + int lit : 24; // lit p size + int *p; // 0 lits +}; + +inline long long hufLength(long long code) { return code & 63; } + +inline long long hufCode(long long code) { return code >> 6; } + +inline void outputBits(int nBits, long long bits, long long &c, int &lc, + char *&out) { + c <<= nBits; + lc += nBits; + + c |= bits; + + while (lc >= 8) *out++ = static_cast((c >> (lc -= 8))); +} + +inline long long getBits(int nBits, long long &c, int &lc, const char *&in) { + while (lc < nBits) { + c = (c << 8) | *(reinterpret_cast(in++)); + lc += 8; + } + + lc -= nBits; + return (c >> lc) & ((1 << nBits) - 1); +} + +// +// ENCODING TABLE BUILDING & (UN)PACKING +// + +// +// Build a "canonical" Huffman code table: +// - for each (uncompressed) symbol, hcode contains the length +// of the corresponding code (in the compressed data) +// - canonical codes are computed and stored in hcode +// - the rules for constructing canonical codes are as follows: +// * shorter codes (if filled with zeroes to the right) +// have a numerically higher value than longer codes +// * for codes with the same length, numerical values +// increase with numerical symbol values +// - because the canonical code table can be constructed from +// symbol lengths alone, the code table can be transmitted +// without sending the actual code values +// - see http://www.compressconsult.com/huffman/ +// + +static void hufCanonicalCodeTable(long long hcode[HUF_ENCSIZE]) { + long long n[59]; + + // + // For each i from 0 through 58, count the + // number of different codes of length i, and + // store the count in n[i]. + // + + for (int i = 0; i <= 58; ++i) n[i] = 0; + + for (int i = 0; i < HUF_ENCSIZE; ++i) n[hcode[i]] += 1; + + // + // For each i from 58 through 1, compute the + // numerically lowest code with length i, and + // store that code in n[i]. + // + + long long c = 0; + + for (int i = 58; i > 0; --i) { + long long nc = ((c + n[i]) >> 1); + n[i] = c; + c = nc; + } + + // + // hcode[i] contains the length, l, of the + // code for symbol i. Assign the next available + // code of length l to the symbol and store both + // l and the code in hcode[i]. + // + + for (int i = 0; i < HUF_ENCSIZE; ++i) { + int l = static_cast(hcode[i]); + + if (l > 0) hcode[i] = l | (n[l]++ << 6); + } +} + +// +// Compute Huffman codes (based on frq input) and store them in frq: +// - code structure is : [63:lsb - 6:msb] | [5-0: bit length]; +// - max code length is 58 bits; +// - codes outside the range [im-iM] have a null length (unused values); +// - original frequencies are destroyed; +// - encoding tables are used by hufEncode() and hufBuildDecTable(); +// + +struct FHeapCompare { + bool operator()(long long *a, long long *b) { return *a > *b; } +}; + +static void hufBuildEncTable( + long long *frq, // io: input frequencies [HUF_ENCSIZE], output table + int *im, // o: min frq index + int *iM) // o: max frq index +{ + // + // This function assumes that when it is called, array frq + // indicates the frequency of all possible symbols in the data + // that are to be Huffman-encoded. (frq[i] contains the number + // of occurrences of symbol i in the data.) + // + // The loop below does three things: + // + // 1) Finds the minimum and maximum indices that point + // to non-zero entries in frq: + // + // frq[im] != 0, and frq[i] == 0 for all i < im + // frq[iM] != 0, and frq[i] == 0 for all i > iM + // + // 2) Fills array fHeap with pointers to all non-zero + // entries in frq. + // + // 3) Initializes array hlink such that hlink[i] == i + // for all array entries. + // + + std::vector hlink(HUF_ENCSIZE); + std::vector fHeap(HUF_ENCSIZE); + + *im = 0; + + while (!frq[*im]) (*im)++; + + int nf = 0; + + for (int i = *im; i < HUF_ENCSIZE; i++) { + hlink[i] = i; + + if (frq[i]) { + fHeap[nf] = &frq[i]; + nf++; + *iM = i; + } + } + + // + // Add a pseudo-symbol, with a frequency count of 1, to frq; + // adjust the fHeap and hlink array accordingly. Function + // hufEncode() uses the pseudo-symbol for run-length encoding. + // + + (*iM)++; + frq[*iM] = 1; + fHeap[nf] = &frq[*iM]; + nf++; + + // + // Build an array, scode, such that scode[i] contains the number + // of bits assigned to symbol i. Conceptually this is done by + // constructing a tree whose leaves are the symbols with non-zero + // frequency: + // + // Make a heap that contains all symbols with a non-zero frequency, + // with the least frequent symbol on top. + // + // Repeat until only one symbol is left on the heap: + // + // Take the two least frequent symbols off the top of the heap. + // Create a new node that has first two nodes as children, and + // whose frequency is the sum of the frequencies of the first + // two nodes. Put the new node back into the heap. + // + // The last node left on the heap is the root of the tree. For each + // leaf node, the distance between the root and the leaf is the length + // of the code for the corresponding symbol. + // + // The loop below doesn't actually build the tree; instead we compute + // the distances of the leaves from the root on the fly. When a new + // node is added to the heap, then that node's descendants are linked + // into a single linear list that starts at the new node, and the code + // lengths of the descendants (that is, their distance from the root + // of the tree) are incremented by one. + // + + std::make_heap(&fHeap[0], &fHeap[nf], FHeapCompare()); + + std::vector scode(HUF_ENCSIZE); + memset(scode.data(), 0, sizeof(long long) * HUF_ENCSIZE); + + while (nf > 1) { + // + // Find the indices, mm and m, of the two smallest non-zero frq + // values in fHeap, add the smallest frq to the second-smallest + // frq, and remove the smallest frq value from fHeap. + // + + int mm = fHeap[0] - frq; + std::pop_heap(&fHeap[0], &fHeap[nf], FHeapCompare()); + --nf; + + int m = fHeap[0] - frq; + std::pop_heap(&fHeap[0], &fHeap[nf], FHeapCompare()); + + frq[m] += frq[mm]; + std::push_heap(&fHeap[0], &fHeap[nf], FHeapCompare()); + + // + // The entries in scode are linked into lists with the + // entries in hlink serving as "next" pointers and with + // the end of a list marked by hlink[j] == j. + // + // Traverse the lists that start at scode[m] and scode[mm]. + // For each element visited, increment the length of the + // corresponding code by one bit. (If we visit scode[j] + // during the traversal, then the code for symbol j becomes + // one bit longer.) + // + // Merge the lists that start at scode[m] and scode[mm] + // into a single list that starts at scode[m]. + // + + // + // Add a bit to all codes in the first list. + // + + for (int j = m;; j = hlink[j]) { + scode[j]++; + + TEXR_ASSERT(scode[j] <= 58); + + if (hlink[j] == j) { + // + // Merge the two lists. + // + + hlink[j] = mm; + break; + } + } + + // + // Add a bit to all codes in the second list + // + + for (int j = mm;; j = hlink[j]) { + scode[j]++; + + TEXR_ASSERT(scode[j] <= 58); + + if (hlink[j] == j) break; + } + } + + // + // Build a canonical Huffman code table, replacing the code + // lengths in scode with (code, code length) pairs. Copy the + // code table from scode into frq. + // + + hufCanonicalCodeTable(scode.data()); + memcpy(frq, scode.data(), sizeof(long long) * HUF_ENCSIZE); +} + +// +// Pack an encoding table: +// - only code lengths, not actual codes, are stored +// - runs of zeroes are compressed as follows: +// +// unpacked packed +// -------------------------------- +// 1 zero 0 (6 bits) +// 2 zeroes 59 +// 3 zeroes 60 +// 4 zeroes 61 +// 5 zeroes 62 +// n zeroes (6 or more) 63 n-6 (6 + 8 bits) +// + +const int SHORT_ZEROCODE_RUN = 59; +const int LONG_ZEROCODE_RUN = 63; +const int SHORTEST_LONG_RUN = 2 + LONG_ZEROCODE_RUN - SHORT_ZEROCODE_RUN; +const int LONGEST_LONG_RUN = 255 + SHORTEST_LONG_RUN; + +static void hufPackEncTable( + const long long *hcode, // i : encoding table [HUF_ENCSIZE] + int im, // i : min hcode index + int iM, // i : max hcode index + char **pcode) // o: ptr to packed table (updated) +{ + char *p = *pcode; + long long c = 0; + int lc = 0; + + for (; im <= iM; im++) { + int l = hufLength(hcode[im]); + + if (l == 0) { + int zerun = 1; + + while ((im < iM) && (zerun < LONGEST_LONG_RUN)) { + if (hufLength(hcode[im + 1]) > 0) break; + im++; + zerun++; + } + + if (zerun >= 2) { + if (zerun >= SHORTEST_LONG_RUN) { + outputBits(6, LONG_ZEROCODE_RUN, c, lc, p); + outputBits(8, zerun - SHORTEST_LONG_RUN, c, lc, p); + } else { + outputBits(6, SHORT_ZEROCODE_RUN + zerun - 2, c, lc, p); + } + continue; + } + } + + outputBits(6, l, c, lc, p); + } + + if (lc > 0) *p++ = (unsigned char)(c << (8 - lc)); + + *pcode = p; +} + +// +// Unpack an encoding table packed by hufPackEncTable(): +// + +static bool hufUnpackEncTable( + const char **pcode, // io: ptr to packed table (updated) + int ni, // i : input size (in bytes) + int im, // i : min hcode index + int iM, // i : max hcode index + long long *hcode) // o: encoding table [HUF_ENCSIZE] +{ + memset(hcode, 0, sizeof(long long) * HUF_ENCSIZE); + + const char *p = *pcode; + long long c = 0; + int lc = 0; + + for (; im <= iM; im++) { + if (p - *pcode >= ni) { + return false; + } + + long long l = hcode[im] = getBits(6, c, lc, p); // code length + + if (l == (long long)LONG_ZEROCODE_RUN) { + if (p - *pcode > ni) { + return false; + } + + int zerun = getBits(8, c, lc, p) + SHORTEST_LONG_RUN; + + if (im + zerun > iM + 1) { + return false; + } + + while (zerun--) hcode[im++] = 0; + + im--; + } else if (l >= (long long)SHORT_ZEROCODE_RUN) { + int zerun = l - SHORT_ZEROCODE_RUN + 2; + + if (im + zerun > iM + 1) { + return false; + } + + while (zerun--) hcode[im++] = 0; + + im--; + } + } + + *pcode = const_cast(p); + + hufCanonicalCodeTable(hcode); + + return true; +} + +// +// DECODING TABLE BUILDING +// + +// +// Clear a newly allocated decoding table so that it contains only zeroes. +// + +static void hufClearDecTable(HufDec *hdecod) // io: (allocated by caller) +// decoding table [HUF_DECSIZE] +{ + for (int i = 0; i < HUF_DECSIZE; i++) { + hdecod[i].len = 0; + hdecod[i].lit = 0; + hdecod[i].p = NULL; + } + // memset(hdecod, 0, sizeof(HufDec) * HUF_DECSIZE); +} + +// +// Build a decoding hash table based on the encoding table hcode: +// - short codes (<= HUF_DECBITS) are resolved with a single table access; +// - long code entry allocations are not optimized, because long codes are +// unfrequent; +// - decoding tables are used by hufDecode(); +// + +static bool hufBuildDecTable(const long long *hcode, // i : encoding table + int im, // i : min index in hcode + int iM, // i : max index in hcode + HufDec *hdecod) // o: (allocated by caller) +// decoding table [HUF_DECSIZE] +{ + // + // Init hashtable & loop on all codes. + // Assumes that hufClearDecTable(hdecod) has already been called. + // + + for (; im <= iM; im++) { + long long c = hufCode(hcode[im]); + int l = hufLength(hcode[im]); + + if (c >> l) { + // + // Error: c is supposed to be an l-bit code, + // but c contains a value that is greater + // than the largest l-bit number. + // + + // invalidTableEntry(); + return false; + } + + if (l > HUF_DECBITS) { + // + // Long code: add a secondary entry + // + + HufDec *pl = hdecod + (c >> (l - HUF_DECBITS)); + + if (pl->len) { + // + // Error: a short code has already + // been stored in table entry *pl. + // + + // invalidTableEntry(); + return false; + } + + pl->lit++; + + if (pl->p) { + int *p = pl->p; + pl->p = new int[pl->lit]; + + for (int i = 0; i < pl->lit - 1; ++i) pl->p[i] = p[i]; + + delete[] p; + } else { + pl->p = new int[1]; + } + + pl->p[pl->lit - 1] = im; + } else if (l) { + // + // Short code: init all primary entries + // + + HufDec *pl = hdecod + (c << (HUF_DECBITS - l)); + + for (long long i = 1ULL << (HUF_DECBITS - l); i > 0; i--, pl++) { + if (pl->len || pl->p) { + // + // Error: a short code or a long code has + // already been stored in table entry *pl. + // + + // invalidTableEntry(); + return false; + } + + pl->len = l; + pl->lit = im; + } + } + } + + return true; +} + +// +// Free the long code entries of a decoding table built by hufBuildDecTable() +// + +static void hufFreeDecTable(HufDec *hdecod) // io: Decoding table +{ + for (int i = 0; i < HUF_DECSIZE; i++) { + if (hdecod[i].p) { + delete[] hdecod[i].p; + hdecod[i].p = 0; + } + } +} + +// +// ENCODING +// + +inline void outputCode(long long code, long long &c, int &lc, char *&out) { + outputBits(hufLength(code), hufCode(code), c, lc, out); +} + +inline void sendCode(long long sCode, int runCount, long long runCode, + long long &c, int &lc, char *&out) { + // + // Output a run of runCount instances of the symbol sCount. + // Output the symbols explicitly, or if that is shorter, output + // the sCode symbol once followed by a runCode symbol and runCount + // expressed as an 8-bit number. + // + + if (hufLength(sCode) + hufLength(runCode) + 8 < hufLength(sCode) * runCount) { + outputCode(sCode, c, lc, out); + outputCode(runCode, c, lc, out); + outputBits(8, runCount, c, lc, out); + } else { + while (runCount-- >= 0) outputCode(sCode, c, lc, out); + } +} + +// +// Encode (compress) ni values based on the Huffman encoding table hcode: +// + +static int hufEncode // return: output size (in bits) + (const long long *hcode, // i : encoding table + const unsigned short *in, // i : uncompressed input buffer + const int ni, // i : input buffer size (in bytes) + int rlc, // i : rl code + char *out) // o: compressed output buffer +{ + char *outStart = out; + long long c = 0; // bits not yet written to out + int lc = 0; // number of valid bits in c (LSB) + int s = in[0]; + int cs = 0; + + // + // Loop on input values + // + + for (int i = 1; i < ni; i++) { + // + // Count same values or send code + // + + if (s == in[i] && cs < 255) { + cs++; + } else { + sendCode(hcode[s], cs, hcode[rlc], c, lc, out); + cs = 0; + } + + s = in[i]; + } + + // + // Send remaining code + // + + sendCode(hcode[s], cs, hcode[rlc], c, lc, out); + + if (lc) *out = (c << (8 - lc)) & 0xff; + + return (out - outStart) * 8 + lc; +} + +// +// DECODING +// + +// +// In order to force the compiler to inline them, +// getChar() and getCode() are implemented as macros +// instead of "inline" functions. +// + +#define getChar(c, lc, in) \ + { \ + c = (c << 8) | *(unsigned char *)(in++); \ + lc += 8; \ + } + +#if 0 +#define getCode(po, rlc, c, lc, in, out, ob, oe) \ + { \ + if (po == rlc) { \ + if (lc < 8) getChar(c, lc, in); \ + \ + lc -= 8; \ + \ + unsigned char cs = (c >> lc); \ + \ + if (out + cs > oe) return false; \ + \ + /* TinyEXR issue 78 */ \ + unsigned short s = out[-1]; \ + \ + while (cs-- > 0) *out++ = s; \ + } else if (out < oe) { \ + *out++ = po; \ + } else { \ + return false; \ + } \ + } +#else +static bool getCode(int po, int rlc, long long &c, int &lc, const char *&in, + const char *in_end, unsigned short *&out, + const unsigned short *ob, const unsigned short *oe) { + (void)ob; + if (po == rlc) { + if (lc < 8) { + /* TinyEXR issue 78 */ + if ((in + 1) >= in_end) { + return false; + } + + getChar(c, lc, in); + } + + lc -= 8; + + unsigned char cs = (c >> lc); + + if (out + cs > oe) return false; + + // Bounds check for safety + // Issue 100. + if ((out - 1) < ob) return false; + unsigned short s = out[-1]; + + while (cs-- > 0) *out++ = s; + } else if (out < oe) { + *out++ = po; + } else { + return false; + } + return true; +} +#endif + +// +// Decode (uncompress) ni bits based on encoding & decoding tables: +// + +static bool hufDecode(const long long *hcode, // i : encoding table + const HufDec *hdecod, // i : decoding table + const char *in, // i : compressed input buffer + int ni, // i : input size (in bits) + int rlc, // i : run-length code + int no, // i : expected output size (in bytes) + unsigned short *out) // o: uncompressed output buffer +{ + long long c = 0; + int lc = 0; + unsigned short *outb = out; // begin + unsigned short *oe = out + no; // end + const char *ie = in + (ni + 7) / 8; // input byte size + + // + // Loop on input bytes + // + + while (in < ie) { + getChar(c, lc, in); + + // + // Access decoding table + // + + while (lc >= HUF_DECBITS) { + const HufDec pl = hdecod[(c >> (lc - HUF_DECBITS)) & HUF_DECMASK]; + + if (pl.len) { + // + // Get short code + // + + lc -= pl.len; + // std::cout << "lit = " << pl.lit << std::endl; + // std::cout << "rlc = " << rlc << std::endl; + // std::cout << "c = " << c << std::endl; + // std::cout << "lc = " << lc << std::endl; + // std::cout << "in = " << in << std::endl; + // std::cout << "out = " << out << std::endl; + // std::cout << "oe = " << oe << std::endl; + if (!getCode(pl.lit, rlc, c, lc, in, ie, out, outb, oe)) { + return false; + } + } else { + if (!pl.p) { + return false; + } + // invalidCode(); // wrong code + + // + // Search long code + // + + int j; + + for (j = 0; j < pl.lit; j++) { + int l = hufLength(hcode[pl.p[j]]); + + while (lc < l && in < ie) // get more bits + getChar(c, lc, in); + + if (lc >= l) { + if (hufCode(hcode[pl.p[j]]) == + ((c >> (lc - l)) & (((long long)(1) << l) - 1))) { + // + // Found : get long code + // + + lc -= l; + if (!getCode(pl.p[j], rlc, c, lc, in, ie, out, outb, oe)) { + return false; + } + break; + } + } + } + + if (j == pl.lit) { + return false; + // invalidCode(); // Not found + } + } + } + } + + // + // Get remaining (short) codes + // + + int i = (8 - ni) & 7; + c >>= i; + lc -= i; + + while (lc > 0) { + const HufDec pl = hdecod[(c << (HUF_DECBITS - lc)) & HUF_DECMASK]; + + if (pl.len) { + lc -= pl.len; + if (!getCode(pl.lit, rlc, c, lc, in, ie, out, outb, oe)) { + return false; + } + } else { + return false; + // invalidCode(); // wrong (long) code + } + } + + if (out - outb != no) { + return false; + } + // notEnoughData (); + + return true; +} + +static void countFrequencies(std::vector &freq, + const unsigned short data[/*n*/], int n) { + for (int i = 0; i < HUF_ENCSIZE; ++i) freq[i] = 0; + + for (int i = 0; i < n; ++i) ++freq[data[i]]; +} + +static void writeUInt(char buf[4], unsigned int i) { + unsigned char *b = (unsigned char *)buf; + + b[0] = i; + b[1] = i >> 8; + b[2] = i >> 16; + b[3] = i >> 24; +} + +static unsigned int readUInt(const char buf[4]) { + const unsigned char *b = (const unsigned char *)buf; + + return (b[0] & 0x000000ff) | ((b[1] << 8) & 0x0000ff00) | + ((b[2] << 16) & 0x00ff0000) | ((b[3] << 24) & 0xff000000); +} + +// +// EXTERNAL INTERFACE +// + +static int hufCompress(const unsigned short raw[], int nRaw, + char compressed[]) { + if (nRaw == 0) return 0; + + std::vector freq(HUF_ENCSIZE); + + countFrequencies(freq, raw, nRaw); + + int im = 0; + int iM = 0; + hufBuildEncTable(freq.data(), &im, &iM); + + char *tableStart = compressed + 20; + char *tableEnd = tableStart; + hufPackEncTable(freq.data(), im, iM, &tableEnd); + int tableLength = tableEnd - tableStart; + + char *dataStart = tableEnd; + int nBits = hufEncode(freq.data(), raw, nRaw, iM, dataStart); + int data_length = (nBits + 7) / 8; + + writeUInt(compressed, im); + writeUInt(compressed + 4, iM); + writeUInt(compressed + 8, tableLength); + writeUInt(compressed + 12, nBits); + writeUInt(compressed + 16, 0); // room for future extensions + + return dataStart + data_length - compressed; +} + +static bool hufUncompress(const char compressed[], int nCompressed, + std::vector *raw) { + if (nCompressed == 0) { + if (raw->size() != 0) return false; + + return false; + } + + int im = readUInt(compressed); + int iM = readUInt(compressed + 4); + // int tableLength = readUInt (compressed + 8); + int nBits = readUInt(compressed + 12); + + if (im < 0 || im >= HUF_ENCSIZE || iM < 0 || iM >= HUF_ENCSIZE) return false; + + const char *ptr = compressed + 20; + + // + // Fast decoder needs at least 2x64-bits of compressed data, and + // needs to be run-able on this platform. Otherwise, fall back + // to the original decoder + // + + // if (FastHufDecoder::enabled() && nBits > 128) + //{ + // FastHufDecoder fhd (ptr, nCompressed - (ptr - compressed), im, iM, iM); + // fhd.decode ((unsigned char*)ptr, nBits, raw, nRaw); + //} + // else + { + std::vector freq(HUF_ENCSIZE); + std::vector hdec(HUF_DECSIZE); + + hufClearDecTable(&hdec.at(0)); + + hufUnpackEncTable(&ptr, nCompressed - (ptr - compressed), im, iM, + &freq.at(0)); + + { + if (nBits > 8 * (nCompressed - (ptr - compressed))) { + return false; + } + + hufBuildDecTable(&freq.at(0), im, iM, &hdec.at(0)); + hufDecode(&freq.at(0), &hdec.at(0), ptr, nBits, iM, raw->size(), + raw->data()); + } + // catch (...) + //{ + // hufFreeDecTable (hdec); + // throw; + //} + + hufFreeDecTable(&hdec.at(0)); + } + + return true; +} + +// +// Functions to compress the range of values in the pixel data +// + +const int USHORT_RANGE = (1 << 16); +const int BITMAP_SIZE = (USHORT_RANGE >> 3); + +static void bitmapFromData(const unsigned short data[/*nData*/], int nData, + unsigned char bitmap[BITMAP_SIZE], + unsigned short &minNonZero, + unsigned short &maxNonZero) { + for (int i = 0; i < BITMAP_SIZE; ++i) bitmap[i] = 0; + + for (int i = 0; i < nData; ++i) bitmap[data[i] >> 3] |= (1 << (data[i] & 7)); + + bitmap[0] &= ~1; // zero is not explicitly stored in + // the bitmap; we assume that the + // data always contain zeroes + minNonZero = BITMAP_SIZE - 1; + maxNonZero = 0; + + for (int i = 0; i < BITMAP_SIZE; ++i) { + if (bitmap[i]) { + if (minNonZero > i) minNonZero = i; + if (maxNonZero < i) maxNonZero = i; + } + } +} + +static unsigned short forwardLutFromBitmap( + const unsigned char bitmap[BITMAP_SIZE], unsigned short lut[USHORT_RANGE]) { + int k = 0; + + for (int i = 0; i < USHORT_RANGE; ++i) { + if ((i == 0) || (bitmap[i >> 3] & (1 << (i & 7)))) + lut[i] = k++; + else + lut[i] = 0; + } + + return k - 1; // maximum value stored in lut[], +} // i.e. number of ones in bitmap minus 1 + +static unsigned short reverseLutFromBitmap( + const unsigned char bitmap[BITMAP_SIZE], unsigned short lut[USHORT_RANGE]) { + int k = 0; + + for (int i = 0; i < USHORT_RANGE; ++i) { + if ((i == 0) || (bitmap[i >> 3] & (1 << (i & 7)))) lut[k++] = i; + } + + int n = k - 1; + + while (k < USHORT_RANGE) lut[k++] = 0; + + return n; // maximum k where lut[k] is non-zero, +} // i.e. number of ones in bitmap minus 1 + +static void applyLut(const unsigned short lut[USHORT_RANGE], + unsigned short data[/*nData*/], int nData) { + for (int i = 0; i < nData; ++i) data[i] = lut[data[i]]; +} + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif // __clang__ + +#ifdef _MSC_VER +#pragma warning(pop) +#endif + +static bool CompressPiz(unsigned char *outPtr, unsigned int *outSize, + const unsigned char *inPtr, size_t inSize, + const std::vector &channelInfo, + int data_width, int num_lines) { + std::vector bitmap(BITMAP_SIZE); + unsigned short minNonZero; + unsigned short maxNonZero; + +#if !MINIZ_LITTLE_ENDIAN + // @todo { PIZ compression on BigEndian architecture. } + TEXR_ASSERT(0); + return false; +#endif + + // Assume `inSize` is multiple of 2 or 4. + std::vector tmpBuffer(inSize / sizeof(unsigned short)); + + std::vector channelData(channelInfo.size()); + unsigned short *tmpBufferEnd = &tmpBuffer.at(0); + + for (size_t c = 0; c < channelData.size(); c++) { + PIZChannelData &cd = channelData[c]; + + cd.start = tmpBufferEnd; + cd.end = cd.start; + + cd.nx = data_width; + cd.ny = num_lines; + // cd.ys = c.channel().ySampling; + + size_t pixelSize = sizeof(int); // UINT and FLOAT + if (channelInfo[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + pixelSize = sizeof(short); + } + + cd.size = static_cast(pixelSize / sizeof(short)); + + tmpBufferEnd += cd.nx * cd.ny * cd.size; + } + + const unsigned char *ptr = inPtr; + for (int y = 0; y < num_lines; ++y) { + for (size_t i = 0; i < channelData.size(); ++i) { + PIZChannelData &cd = channelData[i]; + + // if (modp (y, cd.ys) != 0) + // continue; + + size_t n = static_cast(cd.nx * cd.size); + memcpy(cd.end, ptr, n * sizeof(unsigned short)); + ptr += n * sizeof(unsigned short); + cd.end += n; + } + } + + bitmapFromData(&tmpBuffer.at(0), static_cast(tmpBuffer.size()), + bitmap.data(), minNonZero, maxNonZero); + + std::vector lut(USHORT_RANGE); + unsigned short maxValue = forwardLutFromBitmap(bitmap.data(), lut.data()); + applyLut(lut.data(), &tmpBuffer.at(0), static_cast(tmpBuffer.size())); + + // + // Store range compression info in _outBuffer + // + + char *buf = reinterpret_cast(outPtr); + + memcpy(buf, &minNonZero, sizeof(unsigned short)); + buf += sizeof(unsigned short); + memcpy(buf, &maxNonZero, sizeof(unsigned short)); + buf += sizeof(unsigned short); + + if (minNonZero <= maxNonZero) { + memcpy(buf, reinterpret_cast(&bitmap[0] + minNonZero), + maxNonZero - minNonZero + 1); + buf += maxNonZero - minNonZero + 1; + } + + // + // Apply wavelet encoding + // + + for (size_t i = 0; i < channelData.size(); ++i) { + PIZChannelData &cd = channelData[i]; + + for (int j = 0; j < cd.size; ++j) { + wav2Encode(cd.start + j, cd.nx, cd.size, cd.ny, cd.nx * cd.size, + maxValue); + } + } + + // + // Apply Huffman encoding; append the result to _outBuffer + // + + // length header(4byte), then huff data. Initialize length header with zero, + // then later fill it by `length`. + char *lengthPtr = buf; + int zero = 0; + memcpy(buf, &zero, sizeof(int)); + buf += sizeof(int); + + int length = + hufCompress(&tmpBuffer.at(0), static_cast(tmpBuffer.size()), buf); + memcpy(lengthPtr, &length, sizeof(int)); + + (*outSize) = static_cast( + (reinterpret_cast(buf) - outPtr) + + static_cast(length)); + + // Use uncompressed data when compressed data is larger than uncompressed. + // (Issue 40) + if ((*outSize) >= inSize) { + (*outSize) = static_cast(inSize); + memcpy(outPtr, inPtr, inSize); + } + return true; +} + +static bool DecompressPiz(unsigned char *outPtr, const unsigned char *inPtr, + size_t tmpBufSize, size_t inLen, int num_channels, + const EXRChannelInfo *channels, int data_width, + int num_lines) { + if (inLen == tmpBufSize) { + // Data is not compressed(Issue 40). + memcpy(outPtr, inPtr, inLen); + return true; + } + + std::vector bitmap(BITMAP_SIZE); + unsigned short minNonZero; + unsigned short maxNonZero; + +#if !MINIZ_LITTLE_ENDIAN + // @todo { PIZ compression on BigEndian architecture. } + TEXR_ASSERT(0); + return false; +#endif + + memset(bitmap.data(), 0, BITMAP_SIZE); + + const unsigned char *ptr = inPtr; + // minNonZero = *(reinterpret_cast(ptr)); + tinyexr::cpy2(&minNonZero, reinterpret_cast(ptr)); + // maxNonZero = *(reinterpret_cast(ptr + 2)); + tinyexr::cpy2(&maxNonZero, reinterpret_cast(ptr + 2)); + ptr += 4; + + if (maxNonZero >= BITMAP_SIZE) { + return false; + } + + if (minNonZero <= maxNonZero) { + memcpy(reinterpret_cast(&bitmap[0] + minNonZero), ptr, + maxNonZero - minNonZero + 1); + ptr += maxNonZero - minNonZero + 1; + } + + std::vector lut(USHORT_RANGE); + memset(lut.data(), 0, sizeof(unsigned short) * USHORT_RANGE); + unsigned short maxValue = reverseLutFromBitmap(bitmap.data(), lut.data()); + + // + // Huffman decoding + // + + int length; + + // length = *(reinterpret_cast(ptr)); + tinyexr::cpy4(&length, reinterpret_cast(ptr)); + ptr += sizeof(int); + + if (size_t((ptr - inPtr) + length) > inLen) { + return false; + } + + std::vector tmpBuffer(tmpBufSize); + hufUncompress(reinterpret_cast(ptr), length, &tmpBuffer); + + // + // Wavelet decoding + // + + std::vector channelData(static_cast(num_channels)); + + unsigned short *tmpBufferEnd = &tmpBuffer.at(0); + + for (size_t i = 0; i < static_cast(num_channels); ++i) { + const EXRChannelInfo &chan = channels[i]; + + size_t pixelSize = sizeof(int); // UINT and FLOAT + if (chan.pixel_type == TINYEXR_PIXELTYPE_HALF) { + pixelSize = sizeof(short); + } + + channelData[i].start = tmpBufferEnd; + channelData[i].end = channelData[i].start; + channelData[i].nx = data_width; + channelData[i].ny = num_lines; + // channelData[i].ys = 1; + channelData[i].size = static_cast(pixelSize / sizeof(short)); + + tmpBufferEnd += channelData[i].nx * channelData[i].ny * channelData[i].size; + } + + for (size_t i = 0; i < channelData.size(); ++i) { + PIZChannelData &cd = channelData[i]; + + for (int j = 0; j < cd.size; ++j) { + wav2Decode(cd.start + j, cd.nx, cd.size, cd.ny, cd.nx * cd.size, + maxValue); + } + } + + // + // Expand the pixel data to their original range + // + + applyLut(lut.data(), &tmpBuffer.at(0), static_cast(tmpBufSize)); + + for (int y = 0; y < num_lines; y++) { + for (size_t i = 0; i < channelData.size(); ++i) { + PIZChannelData &cd = channelData[i]; + + // if (modp (y, cd.ys) != 0) + // continue; + + size_t n = static_cast(cd.nx * cd.size); + memcpy(outPtr, cd.end, static_cast(n * sizeof(unsigned short))); + outPtr += n * sizeof(unsigned short); + cd.end += n; + } + } + + return true; +} +#endif // TINYEXR_USE_PIZ + +#if TINYEXR_USE_ZFP +struct ZFPCompressionParam { + double rate; + int precision; + double tolerance; + int type; // TINYEXR_ZFP_COMPRESSIONTYPE_* + + ZFPCompressionParam() { + type = TINYEXR_ZFP_COMPRESSIONTYPE_RATE; + rate = 2.0; + precision = 0; + tolerance = 0.0f; + } +}; + +bool FindZFPCompressionParam(ZFPCompressionParam *param, + const EXRAttribute *attributes, + int num_attributes) { + bool foundType = false; + + for (int i = 0; i < num_attributes; i++) { + if ((strcmp(attributes[i].name, "zfpCompressionType") == 0) && + (attributes[i].size == 1)) { + param->type = static_cast(attributes[i].value[0]); + + foundType = true; + } + } + + if (!foundType) { + return false; + } + + if (param->type == TINYEXR_ZFP_COMPRESSIONTYPE_RATE) { + for (int i = 0; i < num_attributes; i++) { + if ((strcmp(attributes[i].name, "zfpCompressionRate") == 0) && + (attributes[i].size == 8)) { + param->rate = *(reinterpret_cast(attributes[i].value)); + return true; + } + } + } else if (param->type == TINYEXR_ZFP_COMPRESSIONTYPE_PRECISION) { + for (int i = 0; i < num_attributes; i++) { + if ((strcmp(attributes[i].name, "zfpCompressionPrecision") == 0) && + (attributes[i].size == 4)) { + param->rate = *(reinterpret_cast(attributes[i].value)); + return true; + } + } + } else if (param->type == TINYEXR_ZFP_COMPRESSIONTYPE_ACCURACY) { + for (int i = 0; i < num_attributes; i++) { + if ((strcmp(attributes[i].name, "zfpCompressionTolerance") == 0) && + (attributes[i].size == 8)) { + param->tolerance = *(reinterpret_cast(attributes[i].value)); + return true; + } + } + } else { + TEXR_ASSERT(0); + } + + return false; +} + +// Assume pixel format is FLOAT for all channels. +static bool DecompressZfp(float *dst, int dst_width, int dst_num_lines, + int num_channels, const unsigned char *src, + unsigned long src_size, + const ZFPCompressionParam ¶m) { + size_t uncompressed_size = dst_width * dst_num_lines * num_channels; + + if (uncompressed_size == src_size) { + // Data is not compressed(Issue 40). + memcpy(dst, src, src_size); + } + + zfp_stream *zfp = NULL; + zfp_field *field = NULL; + + TEXR_ASSERT((dst_width % 4) == 0); + TEXR_ASSERT((dst_num_lines % 4) == 0); + + if ((dst_width & 3U) || (dst_num_lines & 3U)) { + return false; + } + + field = + zfp_field_2d(reinterpret_cast(const_cast(src)), + zfp_type_float, dst_width, dst_num_lines * num_channels); + zfp = zfp_stream_open(NULL); + + if (param.type == TINYEXR_ZFP_COMPRESSIONTYPE_RATE) { + zfp_stream_set_rate(zfp, param.rate, zfp_type_float, /* dimention */ 2, + /* write random access */ 0); + } else if (param.type == TINYEXR_ZFP_COMPRESSIONTYPE_PRECISION) { + zfp_stream_set_precision(zfp, param.precision, zfp_type_float); + } else if (param.type == TINYEXR_ZFP_COMPRESSIONTYPE_ACCURACY) { + zfp_stream_set_accuracy(zfp, param.tolerance, zfp_type_float); + } else { + TEXR_ASSERT(0); + } + + size_t buf_size = zfp_stream_maximum_size(zfp, field); + std::vector buf(buf_size); + memcpy(&buf.at(0), src, src_size); + + bitstream *stream = stream_open(&buf.at(0), buf_size); + zfp_stream_set_bit_stream(zfp, stream); + zfp_stream_rewind(zfp); + + size_t image_size = dst_width * dst_num_lines; + + for (int c = 0; c < num_channels; c++) { + // decompress 4x4 pixel block. + for (int y = 0; y < dst_num_lines; y += 4) { + for (int x = 0; x < dst_width; x += 4) { + float fblock[16]; + zfp_decode_block_float_2(zfp, fblock); + for (int j = 0; j < 4; j++) { + for (int i = 0; i < 4; i++) { + dst[c * image_size + ((y + j) * dst_width + (x + i))] = + fblock[j * 4 + i]; + } + } + } + } + } + + zfp_field_free(field); + zfp_stream_close(zfp); + stream_close(stream); + + return true; +} + +// Assume pixel format is FLOAT for all channels. +bool CompressZfp(std::vector *outBuf, unsigned int *outSize, + const float *inPtr, int width, int num_lines, int num_channels, + const ZFPCompressionParam ¶m) { + zfp_stream *zfp = NULL; + zfp_field *field = NULL; + + TEXR_ASSERT((width % 4) == 0); + TEXR_ASSERT((num_lines % 4) == 0); + + if ((width & 3U) || (num_lines & 3U)) { + return false; + } + + // create input array. + field = zfp_field_2d(reinterpret_cast(const_cast(inPtr)), + zfp_type_float, width, num_lines * num_channels); + + zfp = zfp_stream_open(NULL); + + if (param.type == TINYEXR_ZFP_COMPRESSIONTYPE_RATE) { + zfp_stream_set_rate(zfp, param.rate, zfp_type_float, 2, 0); + } else if (param.type == TINYEXR_ZFP_COMPRESSIONTYPE_PRECISION) { + zfp_stream_set_precision(zfp, param.precision, zfp_type_float); + } else if (param.type == TINYEXR_ZFP_COMPRESSIONTYPE_ACCURACY) { + zfp_stream_set_accuracy(zfp, param.tolerance, zfp_type_float); + } else { + TEXR_ASSERT(0); + } + + size_t buf_size = zfp_stream_maximum_size(zfp, field); + + outBuf->resize(buf_size); + + bitstream *stream = stream_open(&outBuf->at(0), buf_size); + zfp_stream_set_bit_stream(zfp, stream); + zfp_field_free(field); + + size_t image_size = width * num_lines; + + for (int c = 0; c < num_channels; c++) { + // compress 4x4 pixel block. + for (int y = 0; y < num_lines; y += 4) { + for (int x = 0; x < width; x += 4) { + float fblock[16]; + for (int j = 0; j < 4; j++) { + for (int i = 0; i < 4; i++) { + fblock[j * 4 + i] = + inPtr[c * image_size + ((y + j) * width + (x + i))]; + } + } + zfp_encode_block_float_2(zfp, fblock); + } + } + } + + zfp_stream_flush(zfp); + (*outSize) = zfp_stream_compressed_size(zfp); + + zfp_stream_close(zfp); + + return true; +} + +#endif + +// +// ----------------------------------------------------------------- +// + +// TODO(syoyo): Refactor function arguments. +static bool DecodePixelData(/* out */ unsigned char **out_images, + const int *requested_pixel_types, + const unsigned char *data_ptr, size_t data_len, + int compression_type, int line_order, int width, + int height, int x_stride, int y, int line_no, + int num_lines, size_t pixel_data_size, + size_t num_attributes, + const EXRAttribute *attributes, size_t num_channels, + const EXRChannelInfo *channels, + const std::vector &channel_offset_list) { + if (compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { // PIZ +#if TINYEXR_USE_PIZ + if ((width == 0) || (num_lines == 0) || (pixel_data_size == 0)) { + // Invalid input #90 + return false; + } + + // Allocate original data size. + std::vector outBuf(static_cast( + static_cast(width * num_lines) * pixel_data_size)); + size_t tmpBufLen = outBuf.size(); + + bool ret = tinyexr::DecompressPiz( + reinterpret_cast(&outBuf.at(0)), data_ptr, tmpBufLen, + data_len, static_cast(num_channels), channels, width, num_lines); + + if (!ret) { + return false; + } + + // For PIZ_COMPRESSION: + // pixel sample data for channel 0 for scanline 0 + // pixel sample data for channel 1 for scanline 0 + // pixel sample data for channel ... for scanline 0 + // pixel sample data for channel n for scanline 0 + // pixel sample data for channel 0 for scanline 1 + // pixel sample data for channel 1 for scanline 1 + // pixel sample data for channel ... for scanline 1 + // pixel sample data for channel n for scanline 1 + // ... + for (size_t c = 0; c < static_cast(num_channels); c++) { + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + for (size_t v = 0; v < static_cast(num_lines); v++) { + const unsigned short *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + FP16 hf; + + // hf.u = line_ptr[u]; + // use `cpy` to avoid unaligned memory access when compiler's + // optimization is on. + tinyexr::cpy2(&(hf.u), line_ptr + u); + + tinyexr::swap2(reinterpret_cast(&hf.u)); + + if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + unsigned short *image = + reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += static_cast( + (height - 1 - (line_no + static_cast(v)))) * + static_cast(x_stride) + + u; + } + *image = hf.u; + } else { // HALF -> FLOAT + FP32 f32 = half_to_float(hf); + float *image = reinterpret_cast(out_images)[c]; + size_t offset = 0; + if (line_order == 0) { + offset = (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + offset = static_cast( + (height - 1 - (line_no + static_cast(v)))) * + static_cast(x_stride) + + u; + } + image += offset; + *image = f32.f; + } + } + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_UINT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_UINT); + + for (size_t v = 0; v < static_cast(num_lines); v++) { + const unsigned int *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + unsigned int val; + // val = line_ptr[u]; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(&val); + + unsigned int *image = + reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += static_cast( + (height - 1 - (line_no + static_cast(v)))) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT); + for (size_t v = 0; v < static_cast(num_lines); v++) { + const float *line_ptr = reinterpret_cast(&outBuf.at( + v * pixel_data_size * static_cast(x_stride) + + channel_offset_list[c] * static_cast(x_stride))); + for (size_t u = 0; u < static_cast(width); u++) { + float val; + // val = line_ptr[u]; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(reinterpret_cast(&val)); + + float *image = reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += static_cast( + (height - 1 - (line_no + static_cast(v)))) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else { + TEXR_ASSERT(0); + } + } +#else + TEXR_ASSERT(0 && "PIZ is enabled in this build"); + return false; +#endif + + } else if (compression_type == TINYEXR_COMPRESSIONTYPE_ZIPS || + compression_type == TINYEXR_COMPRESSIONTYPE_ZIP) { + // Allocate original data size. + std::vector outBuf(static_cast(width) * + static_cast(num_lines) * + pixel_data_size); + + unsigned long dstLen = static_cast(outBuf.size()); + TEXR_ASSERT(dstLen > 0); + if (!tinyexr::DecompressZip( + reinterpret_cast(&outBuf.at(0)), &dstLen, data_ptr, + static_cast(data_len))) { + return false; + } + + // For ZIP_COMPRESSION: + // pixel sample data for channel 0 for scanline 0 + // pixel sample data for channel 1 for scanline 0 + // pixel sample data for channel ... for scanline 0 + // pixel sample data for channel n for scanline 0 + // pixel sample data for channel 0 for scanline 1 + // pixel sample data for channel 1 for scanline 1 + // pixel sample data for channel ... for scanline 1 + // pixel sample data for channel n for scanline 1 + // ... + for (size_t c = 0; c < static_cast(num_channels); c++) { + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + for (size_t v = 0; v < static_cast(num_lines); v++) { + const unsigned short *line_ptr = reinterpret_cast( + &outBuf.at(v * static_cast(pixel_data_size) * + static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + tinyexr::FP16 hf; + + // hf.u = line_ptr[u]; + tinyexr::cpy2(&(hf.u), line_ptr + u); + + tinyexr::swap2(reinterpret_cast(&hf.u)); + + if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + unsigned short *image = + reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = hf.u; + } else { // HALF -> FLOAT + tinyexr::FP32 f32 = half_to_float(hf); + float *image = reinterpret_cast(out_images)[c]; + size_t offset = 0; + if (line_order == 0) { + offset = (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + offset = (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + image += offset; + + *image = f32.f; + } + } + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_UINT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_UINT); + + for (size_t v = 0; v < static_cast(num_lines); v++) { + const unsigned int *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + unsigned int val; + // val = line_ptr[u]; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(&val); + + unsigned int *image = + reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT); + for (size_t v = 0; v < static_cast(num_lines); v++) { + const float *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + float val; + // val = line_ptr[u]; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(reinterpret_cast(&val)); + + float *image = reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else { + TEXR_ASSERT(0); + return false; + } + } + } else if (compression_type == TINYEXR_COMPRESSIONTYPE_RLE) { + // Allocate original data size. + std::vector outBuf(static_cast(width) * + static_cast(num_lines) * + pixel_data_size); + + unsigned long dstLen = static_cast(outBuf.size()); + if (dstLen == 0) { + return false; + } + + if (!tinyexr::DecompressRle( + reinterpret_cast(&outBuf.at(0)), dstLen, data_ptr, + static_cast(data_len))) { + return false; + } + + // For RLE_COMPRESSION: + // pixel sample data for channel 0 for scanline 0 + // pixel sample data for channel 1 for scanline 0 + // pixel sample data for channel ... for scanline 0 + // pixel sample data for channel n for scanline 0 + // pixel sample data for channel 0 for scanline 1 + // pixel sample data for channel 1 for scanline 1 + // pixel sample data for channel ... for scanline 1 + // pixel sample data for channel n for scanline 1 + // ... + for (size_t c = 0; c < static_cast(num_channels); c++) { + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + for (size_t v = 0; v < static_cast(num_lines); v++) { + const unsigned short *line_ptr = reinterpret_cast( + &outBuf.at(v * static_cast(pixel_data_size) * + static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + tinyexr::FP16 hf; + + // hf.u = line_ptr[u]; + tinyexr::cpy2(&(hf.u), line_ptr + u); + + tinyexr::swap2(reinterpret_cast(&hf.u)); + + if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + unsigned short *image = + reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = hf.u; + } else { // HALF -> FLOAT + tinyexr::FP32 f32 = half_to_float(hf); + float *image = reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = f32.f; + } + } + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_UINT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_UINT); + + for (size_t v = 0; v < static_cast(num_lines); v++) { + const unsigned int *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + unsigned int val; + // val = line_ptr[u]; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(&val); + + unsigned int *image = + reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT); + for (size_t v = 0; v < static_cast(num_lines); v++) { + const float *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + float val; + // val = line_ptr[u]; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(reinterpret_cast(&val)); + + float *image = reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else { + TEXR_ASSERT(0); + return false; + } + } + } else if (compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { +#if TINYEXR_USE_ZFP + tinyexr::ZFPCompressionParam zfp_compression_param; + if (!FindZFPCompressionParam(&zfp_compression_param, attributes, + num_attributes)) { + TEXR_ASSERT(0); + return false; + } + + // Allocate original data size. + std::vector outBuf(static_cast(width) * + static_cast(num_lines) * + pixel_data_size); + + unsigned long dstLen = outBuf.size(); + TEXR_ASSERT(dstLen > 0); + tinyexr::DecompressZfp(reinterpret_cast(&outBuf.at(0)), width, + num_lines, num_channels, data_ptr, + static_cast(data_len), + zfp_compression_param); + + // For ZFP_COMPRESSION: + // pixel sample data for channel 0 for scanline 0 + // pixel sample data for channel 1 for scanline 0 + // pixel sample data for channel ... for scanline 0 + // pixel sample data for channel n for scanline 0 + // pixel sample data for channel 0 for scanline 1 + // pixel sample data for channel 1 for scanline 1 + // pixel sample data for channel ... for scanline 1 + // pixel sample data for channel n for scanline 1 + // ... + for (size_t c = 0; c < static_cast(num_channels); c++) { + TEXR_ASSERT(channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT); + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + TEXR_ASSERT(requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT); + for (size_t v = 0; v < static_cast(num_lines); v++) { + const float *line_ptr = reinterpret_cast( + &outBuf.at(v * pixel_data_size * static_cast(width) + + channel_offset_list[c] * static_cast(width))); + for (size_t u = 0; u < static_cast(width); u++) { + float val; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(reinterpret_cast(&val)); + + float *image = reinterpret_cast(out_images)[c]; + if (line_order == 0) { + image += (static_cast(line_no) + v) * + static_cast(x_stride) + + u; + } else { + image += (static_cast(height) - 1U - + (static_cast(line_no) + v)) * + static_cast(x_stride) + + u; + } + *image = val; + } + } + } else { + TEXR_ASSERT(0); + return false; + } + } +#else + (void)attributes; + (void)num_attributes; + (void)num_channels; + TEXR_ASSERT(0); + return false; +#endif + } else if (compression_type == TINYEXR_COMPRESSIONTYPE_NONE) { + for (size_t c = 0; c < num_channels; c++) { + for (size_t v = 0; v < static_cast(num_lines); v++) { + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + const unsigned short *line_ptr = + reinterpret_cast( + data_ptr + v * pixel_data_size * size_t(width) + + channel_offset_list[c] * static_cast(width)); + + if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + unsigned short *outLine = + reinterpret_cast(out_images[c]); + if (line_order == 0) { + outLine += (size_t(y) + v) * size_t(x_stride); + } else { + outLine += + (size_t(height) - 1 - (size_t(y) + v)) * size_t(x_stride); + } + + for (int u = 0; u < width; u++) { + tinyexr::FP16 hf; + + // hf.u = line_ptr[u]; + tinyexr::cpy2(&(hf.u), line_ptr + u); + + tinyexr::swap2(reinterpret_cast(&hf.u)); + + outLine[u] = hf.u; + } + } else if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT) { + float *outLine = reinterpret_cast(out_images[c]); + if (line_order == 0) { + outLine += (size_t(y) + v) * size_t(x_stride); + } else { + outLine += + (size_t(height) - 1 - (size_t(y) + v)) * size_t(x_stride); + } + + if (reinterpret_cast(line_ptr + width) > + (data_ptr + data_len)) { + // Insufficient data size + return false; + } + + for (int u = 0; u < width; u++) { + tinyexr::FP16 hf; + + // address may not be aliged. use byte-wise copy for safety.#76 + // hf.u = line_ptr[u]; + tinyexr::cpy2(&(hf.u), line_ptr + u); + + tinyexr::swap2(reinterpret_cast(&hf.u)); + + tinyexr::FP32 f32 = half_to_float(hf); + + outLine[u] = f32.f; + } + } else { + TEXR_ASSERT(0); + return false; + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + const float *line_ptr = reinterpret_cast( + data_ptr + v * pixel_data_size * size_t(width) + + channel_offset_list[c] * static_cast(width)); + + float *outLine = reinterpret_cast(out_images[c]); + if (line_order == 0) { + outLine += (size_t(y) + v) * size_t(x_stride); + } else { + outLine += + (size_t(height) - 1 - (size_t(y) + v)) * size_t(x_stride); + } + + if (reinterpret_cast(line_ptr + width) > + (data_ptr + data_len)) { + // Insufficient data size + return false; + } + + for (int u = 0; u < width; u++) { + float val; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(reinterpret_cast(&val)); + + outLine[u] = val; + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_UINT) { + const unsigned int *line_ptr = reinterpret_cast( + data_ptr + v * pixel_data_size * size_t(width) + + channel_offset_list[c] * static_cast(width)); + + unsigned int *outLine = + reinterpret_cast(out_images[c]); + if (line_order == 0) { + outLine += (size_t(y) + v) * size_t(x_stride); + } else { + outLine += + (size_t(height) - 1 - (size_t(y) + v)) * size_t(x_stride); + } + + for (int u = 0; u < width; u++) { + if (reinterpret_cast(line_ptr + u) >= + (data_ptr + data_len)) { + // Corrupsed data? + return false; + } + + unsigned int val; + tinyexr::cpy4(&val, line_ptr + u); + + tinyexr::swap4(reinterpret_cast(&val)); + + outLine[u] = val; + } + } + } + } + } + + return true; +} + +static bool DecodeTiledPixelData( + unsigned char **out_images, int *width, int *height, + const int *requested_pixel_types, const unsigned char *data_ptr, + size_t data_len, int compression_type, int line_order, int data_width, + int data_height, int tile_offset_x, int tile_offset_y, int tile_size_x, + int tile_size_y, size_t pixel_data_size, size_t num_attributes, + const EXRAttribute *attributes, size_t num_channels, + const EXRChannelInfo *channels, + const std::vector &channel_offset_list) { + TEXR_ASSERT(tile_offset_x * tile_size_x < data_width); + TEXR_ASSERT(tile_offset_y * tile_size_y < data_height); + + // Compute actual image size in a tile. + if ((tile_offset_x + 1) * tile_size_x >= data_width) { + (*width) = data_width - (tile_offset_x * tile_size_x); + } else { + (*width) = tile_size_x; + } + + if ((tile_offset_y + 1) * tile_size_y >= data_height) { + (*height) = data_height - (tile_offset_y * tile_size_y); + } else { + (*height) = tile_size_y; + } + + // Image size = tile size. + return DecodePixelData(out_images, requested_pixel_types, data_ptr, data_len, + compression_type, line_order, (*width), tile_size_y, + /* stride */ tile_size_x, /* y */ 0, /* line_no */ 0, + (*height), pixel_data_size, num_attributes, attributes, + num_channels, channels, channel_offset_list); +} + +static bool ComputeChannelLayout(std::vector *channel_offset_list, + int *pixel_data_size, size_t *channel_offset, + int num_channels, + const EXRChannelInfo *channels) { + channel_offset_list->resize(static_cast(num_channels)); + + (*pixel_data_size) = 0; + (*channel_offset) = 0; + + for (size_t c = 0; c < static_cast(num_channels); c++) { + (*channel_offset_list)[c] = (*channel_offset); + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + (*pixel_data_size) += sizeof(unsigned short); + (*channel_offset) += sizeof(unsigned short); + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + (*pixel_data_size) += sizeof(float); + (*channel_offset) += sizeof(float); + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_UINT) { + (*pixel_data_size) += sizeof(unsigned int); + (*channel_offset) += sizeof(unsigned int); + } else { + // ??? + return false; + } + } + return true; +} + +static unsigned char **AllocateImage(int num_channels, + const EXRChannelInfo *channels, + const int *requested_pixel_types, + int data_width, int data_height) { + unsigned char **images = + reinterpret_cast(static_cast( + malloc(sizeof(float *) * static_cast(num_channels)))); + + for (size_t c = 0; c < static_cast(num_channels); c++) { + size_t data_len = + static_cast(data_width) * static_cast(data_height); + if (channels[c].pixel_type == TINYEXR_PIXELTYPE_HALF) { + // pixel_data_size += sizeof(unsigned short); + // channel_offset += sizeof(unsigned short); + // Alloc internal image for half type. + if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + images[c] = + reinterpret_cast(static_cast( + malloc(sizeof(unsigned short) * data_len))); + } else if (requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT) { + images[c] = reinterpret_cast( + static_cast(malloc(sizeof(float) * data_len))); + } else { + TEXR_ASSERT(0); + } + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_FLOAT) { + // pixel_data_size += sizeof(float); + // channel_offset += sizeof(float); + images[c] = reinterpret_cast( + static_cast(malloc(sizeof(float) * data_len))); + } else if (channels[c].pixel_type == TINYEXR_PIXELTYPE_UINT) { + // pixel_data_size += sizeof(unsigned int); + // channel_offset += sizeof(unsigned int); + images[c] = reinterpret_cast( + static_cast(malloc(sizeof(unsigned int) * data_len))); + } else { + TEXR_ASSERT(0); + } + } + + return images; +} + +static int ParseEXRHeader(HeaderInfo *info, bool *empty_header, + const EXRVersion *version, std::string *err, + const unsigned char *buf, size_t size) { + const char *marker = reinterpret_cast(&buf[0]); + + if (empty_header) { + (*empty_header) = false; + } + + if (version->multipart) { + if (size > 0 && marker[0] == '\0') { + // End of header list. + if (empty_header) { + (*empty_header) = true; + } + return TINYEXR_SUCCESS; + } + } + + // According to the spec, the header of every OpenEXR file must contain at + // least the following attributes: + // + // channels chlist + // compression compression + // dataWindow box2i + // displayWindow box2i + // lineOrder lineOrder + // pixelAspectRatio float + // screenWindowCenter v2f + // screenWindowWidth float + bool has_channels = false; + bool has_compression = false; + bool has_data_window = false; + bool has_display_window = false; + bool has_line_order = false; + bool has_pixel_aspect_ratio = false; + bool has_screen_window_center = false; + bool has_screen_window_width = false; + + info->data_window[0] = 0; + info->data_window[1] = 0; + info->data_window[2] = 0; + info->data_window[3] = 0; + info->line_order = 0; // @fixme + info->display_window[0] = 0; + info->display_window[1] = 0; + info->display_window[2] = 0; + info->display_window[3] = 0; + info->screen_window_center[0] = 0.0f; + info->screen_window_center[1] = 0.0f; + info->screen_window_width = -1.0f; + info->pixel_aspect_ratio = -1.0f; + + info->tile_size_x = -1; + info->tile_size_y = -1; + info->tile_level_mode = -1; + info->tile_rounding_mode = -1; + + info->attributes.clear(); + + // Read attributes + size_t orig_size = size; + for (size_t nattr = 0; nattr < TINYEXR_MAX_HEADER_ATTRIBUTES; nattr++) { + if (0 == size) { + if (err) { + (*err) += "Insufficient data size for attributes.\n"; + } + return TINYEXR_ERROR_INVALID_DATA; + } else if (marker[0] == '\0') { + size--; + break; + } + + std::string attr_name; + std::string attr_type; + std::vector data; + size_t marker_size; + if (!tinyexr::ReadAttribute(&attr_name, &attr_type, &data, &marker_size, + marker, size)) { + if (err) { + (*err) += "Failed to read attribute.\n"; + } + return TINYEXR_ERROR_INVALID_DATA; + } + marker += marker_size; + size -= marker_size; + + if (version->tiled && attr_name.compare("tiles") == 0) { + unsigned int x_size, y_size; + unsigned char tile_mode; + TEXR_ASSERT(data.size() == 9); + memcpy(&x_size, &data.at(0), sizeof(int)); + memcpy(&y_size, &data.at(4), sizeof(int)); + tile_mode = data[8]; + tinyexr::swap4(&x_size); + tinyexr::swap4(&y_size); + + info->tile_size_x = static_cast(x_size); + info->tile_size_y = static_cast(y_size); + + // mode = levelMode + roundingMode * 16 + info->tile_level_mode = tile_mode & 0x3; + info->tile_rounding_mode = (tile_mode >> 4) & 0x1; + + } else if (attr_name.compare("compression") == 0) { + bool ok = false; + if (data[0] < TINYEXR_COMPRESSIONTYPE_PIZ) { + ok = true; + } + + if (data[0] == TINYEXR_COMPRESSIONTYPE_PIZ) { +#if TINYEXR_USE_PIZ + ok = true; +#else + if (err) { + (*err) = "PIZ compression is not supported."; + } + return TINYEXR_ERROR_UNSUPPORTED_FORMAT; +#endif + } + + if (data[0] == TINYEXR_COMPRESSIONTYPE_ZFP) { +#if TINYEXR_USE_ZFP + ok = true; +#else + if (err) { + (*err) = "ZFP compression is not supported."; + } + return TINYEXR_ERROR_UNSUPPORTED_FORMAT; +#endif + } + + if (!ok) { + if (err) { + (*err) = "Unknown compression type."; + } + return TINYEXR_ERROR_UNSUPPORTED_FORMAT; + } + + info->compression_type = static_cast(data[0]); + has_compression = true; + + } else if (attr_name.compare("channels") == 0) { + // name: zero-terminated string, from 1 to 255 bytes long + // pixel type: int, possible values are: UINT = 0 HALF = 1 FLOAT = 2 + // pLinear: unsigned char, possible values are 0 and 1 + // reserved: three chars, should be zero + // xSampling: int + // ySampling: int + + if (!ReadChannelInfo(info->channels, data)) { + if (err) { + (*err) += "Failed to parse channel info.\n"; + } + return TINYEXR_ERROR_INVALID_DATA; + } + + if (info->channels.size() < 1) { + if (err) { + (*err) += "# of channels is zero.\n"; + } + return TINYEXR_ERROR_INVALID_DATA; + } + + has_channels = true; + + } else if (attr_name.compare("dataWindow") == 0) { + if (data.size() >= 16) { + memcpy(&info->data_window[0], &data.at(0), sizeof(int)); + memcpy(&info->data_window[1], &data.at(4), sizeof(int)); + memcpy(&info->data_window[2], &data.at(8), sizeof(int)); + memcpy(&info->data_window[3], &data.at(12), sizeof(int)); + tinyexr::swap4(reinterpret_cast(&info->data_window[0])); + tinyexr::swap4(reinterpret_cast(&info->data_window[1])); + tinyexr::swap4(reinterpret_cast(&info->data_window[2])); + tinyexr::swap4(reinterpret_cast(&info->data_window[3])); + has_data_window = true; + } + } else if (attr_name.compare("displayWindow") == 0) { + if (data.size() >= 16) { + memcpy(&info->display_window[0], &data.at(0), sizeof(int)); + memcpy(&info->display_window[1], &data.at(4), sizeof(int)); + memcpy(&info->display_window[2], &data.at(8), sizeof(int)); + memcpy(&info->display_window[3], &data.at(12), sizeof(int)); + tinyexr::swap4( + reinterpret_cast(&info->display_window[0])); + tinyexr::swap4( + reinterpret_cast(&info->display_window[1])); + tinyexr::swap4( + reinterpret_cast(&info->display_window[2])); + tinyexr::swap4( + reinterpret_cast(&info->display_window[3])); + + has_display_window = true; + } + } else if (attr_name.compare("lineOrder") == 0) { + if (data.size() >= 1) { + info->line_order = static_cast(data[0]); + has_line_order = true; + } + } else if (attr_name.compare("pixelAspectRatio") == 0) { + if (data.size() >= sizeof(float)) { + memcpy(&info->pixel_aspect_ratio, &data.at(0), sizeof(float)); + tinyexr::swap4( + reinterpret_cast(&info->pixel_aspect_ratio)); + has_pixel_aspect_ratio = true; + } + } else if (attr_name.compare("screenWindowCenter") == 0) { + if (data.size() >= 8) { + memcpy(&info->screen_window_center[0], &data.at(0), sizeof(float)); + memcpy(&info->screen_window_center[1], &data.at(4), sizeof(float)); + tinyexr::swap4( + reinterpret_cast(&info->screen_window_center[0])); + tinyexr::swap4( + reinterpret_cast(&info->screen_window_center[1])); + has_screen_window_center = true; + } + } else if (attr_name.compare("screenWindowWidth") == 0) { + if (data.size() >= sizeof(float)) { + memcpy(&info->screen_window_width, &data.at(0), sizeof(float)); + tinyexr::swap4( + reinterpret_cast(&info->screen_window_width)); + + has_screen_window_width = true; + } + } else if (attr_name.compare("chunkCount") == 0) { + if (data.size() >= sizeof(int)) { + memcpy(&info->chunk_count, &data.at(0), sizeof(int)); + tinyexr::swap4(reinterpret_cast(&info->chunk_count)); + } + } else { + // Custom attribute(up to TINYEXR_MAX_CUSTOM_ATTRIBUTES) + if (info->attributes.size() < TINYEXR_MAX_CUSTOM_ATTRIBUTES) { + EXRAttribute attrib; +#ifdef _MSC_VER + strncpy_s(attrib.name, attr_name.c_str(), 255); + strncpy_s(attrib.type, attr_type.c_str(), 255); +#else + strncpy(attrib.name, attr_name.c_str(), 255); + strncpy(attrib.type, attr_type.c_str(), 255); +#endif + attrib.name[255] = '\0'; + attrib.type[255] = '\0'; + attrib.size = static_cast(data.size()); + attrib.value = static_cast(malloc(data.size())); + memcpy(reinterpret_cast(attrib.value), &data.at(0), + data.size()); + info->attributes.push_back(attrib); + } + } + } + + // Check if required attributes exist + { + std::stringstream ss_err; + + if (!has_compression) { + ss_err << "\"compression\" attribute not found in the header." + << std::endl; + } + + if (!has_channels) { + ss_err << "\"channels\" attribute not found in the header." << std::endl; + } + + if (!has_line_order) { + ss_err << "\"lineOrder\" attribute not found in the header." << std::endl; + } + + if (!has_display_window) { + ss_err << "\"displayWindow\" attribute not found in the header." + << std::endl; + } + + if (!has_data_window) { + ss_err << "\"dataWindow\" attribute not found in the header or invalid." + << std::endl; + } + + if (!has_pixel_aspect_ratio) { + ss_err << "\"pixelAspectRatio\" attribute not found in the header." + << std::endl; + } + + if (!has_screen_window_width) { + ss_err << "\"screenWindowWidth\" attribute not found in the header." + << std::endl; + } + + if (!has_screen_window_center) { + ss_err << "\"screenWindowCenter\" attribute not found in the header." + << std::endl; + } + + if (!(ss_err.str().empty())) { + if (err) { + (*err) += ss_err.str(); + } + return TINYEXR_ERROR_INVALID_HEADER; + } + } + + info->header_len = static_cast(orig_size - size); + + return TINYEXR_SUCCESS; +} + +// C++ HeaderInfo to C EXRHeader conversion. +static void ConvertHeader(EXRHeader *exr_header, const HeaderInfo &info) { + exr_header->pixel_aspect_ratio = info.pixel_aspect_ratio; + exr_header->screen_window_center[0] = info.screen_window_center[0]; + exr_header->screen_window_center[1] = info.screen_window_center[1]; + exr_header->screen_window_width = info.screen_window_width; + exr_header->chunk_count = info.chunk_count; + exr_header->display_window[0] = info.display_window[0]; + exr_header->display_window[1] = info.display_window[1]; + exr_header->display_window[2] = info.display_window[2]; + exr_header->display_window[3] = info.display_window[3]; + exr_header->data_window[0] = info.data_window[0]; + exr_header->data_window[1] = info.data_window[1]; + exr_header->data_window[2] = info.data_window[2]; + exr_header->data_window[3] = info.data_window[3]; + exr_header->line_order = info.line_order; + exr_header->compression_type = info.compression_type; + + exr_header->tile_size_x = info.tile_size_x; + exr_header->tile_size_y = info.tile_size_y; + exr_header->tile_level_mode = info.tile_level_mode; + exr_header->tile_rounding_mode = info.tile_rounding_mode; + + exr_header->num_channels = static_cast(info.channels.size()); + + exr_header->channels = static_cast(malloc( + sizeof(EXRChannelInfo) * static_cast(exr_header->num_channels))); + for (size_t c = 0; c < static_cast(exr_header->num_channels); c++) { +#ifdef _MSC_VER + strncpy_s(exr_header->channels[c].name, info.channels[c].name.c_str(), 255); +#else + strncpy(exr_header->channels[c].name, info.channels[c].name.c_str(), 255); +#endif + // manually add '\0' for safety. + exr_header->channels[c].name[255] = '\0'; + + exr_header->channels[c].pixel_type = info.channels[c].pixel_type; + exr_header->channels[c].p_linear = info.channels[c].p_linear; + exr_header->channels[c].x_sampling = info.channels[c].x_sampling; + exr_header->channels[c].y_sampling = info.channels[c].y_sampling; + } + + exr_header->pixel_types = static_cast( + malloc(sizeof(int) * static_cast(exr_header->num_channels))); + for (size_t c = 0; c < static_cast(exr_header->num_channels); c++) { + exr_header->pixel_types[c] = info.channels[c].pixel_type; + } + + // Initially fill with values of `pixel_types` + exr_header->requested_pixel_types = static_cast( + malloc(sizeof(int) * static_cast(exr_header->num_channels))); + for (size_t c = 0; c < static_cast(exr_header->num_channels); c++) { + exr_header->requested_pixel_types[c] = info.channels[c].pixel_type; + } + + exr_header->num_custom_attributes = static_cast(info.attributes.size()); + + if (exr_header->num_custom_attributes > 0) { + // TODO(syoyo): Report warning when # of attributes exceeds + // `TINYEXR_MAX_CUSTOM_ATTRIBUTES` + if (exr_header->num_custom_attributes > TINYEXR_MAX_CUSTOM_ATTRIBUTES) { + exr_header->num_custom_attributes = TINYEXR_MAX_CUSTOM_ATTRIBUTES; + } + + exr_header->custom_attributes = static_cast(malloc( + sizeof(EXRAttribute) * size_t(exr_header->num_custom_attributes))); + + for (size_t i = 0; i < info.attributes.size(); i++) { + memcpy(exr_header->custom_attributes[i].name, info.attributes[i].name, + 256); + memcpy(exr_header->custom_attributes[i].type, info.attributes[i].type, + 256); + exr_header->custom_attributes[i].size = info.attributes[i].size; + // Just copy poiner + exr_header->custom_attributes[i].value = info.attributes[i].value; + } + + } else { + exr_header->custom_attributes = NULL; + } + + exr_header->header_len = info.header_len; +} + +static int DecodeChunk(EXRImage *exr_image, const EXRHeader *exr_header, + const std::vector &offsets, + const unsigned char *head, const size_t size, + std::string *err) { + int num_channels = exr_header->num_channels; + + int num_scanline_blocks = 1; + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZIP) { + num_scanline_blocks = 16; + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { + num_scanline_blocks = 32; + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { + num_scanline_blocks = 16; + } + + int data_width = exr_header->data_window[2] - exr_header->data_window[0] + 1; + int data_height = exr_header->data_window[3] - exr_header->data_window[1] + 1; + + if ((data_width < 0) || (data_height < 0)) { + if (err) { + std::stringstream ss; + ss << "Invalid data width or data height: " << data_width << ", " + << data_height << std::endl; + (*err) += ss.str(); + } + return TINYEXR_ERROR_INVALID_DATA; + } + + // Do not allow too large data_width and data_height. header invalid? + { + const int threshold = 1024 * 8192; // heuristics + if ((data_width > threshold) || (data_height > threshold)) { + if (err) { + std::stringstream ss; + ss << "data_with or data_height too large. data_width: " << data_width + << ", " + << "data_height = " << data_height << std::endl; + (*err) += ss.str(); + } + return TINYEXR_ERROR_INVALID_DATA; + } + } + + size_t num_blocks = offsets.size(); + + std::vector channel_offset_list; + int pixel_data_size = 0; + size_t channel_offset = 0; + if (!tinyexr::ComputeChannelLayout(&channel_offset_list, &pixel_data_size, + &channel_offset, num_channels, + exr_header->channels)) { + if (err) { + (*err) += "Failed to compute channel layout.\n"; + } + return TINYEXR_ERROR_INVALID_DATA; + } + + bool invalid_data = false; // TODO(LTE): Use atomic lock for MT safety. + + if (exr_header->tiled) { + // value check + if (exr_header->tile_size_x < 0) { + if (err) { + std::stringstream ss; + ss << "Invalid tile size x : " << exr_header->tile_size_x << "\n"; + (*err) += ss.str(); + } + return TINYEXR_ERROR_INVALID_HEADER; + } + + if (exr_header->tile_size_y < 0) { + if (err) { + std::stringstream ss; + ss << "Invalid tile size y : " << exr_header->tile_size_y << "\n"; + (*err) += ss.str(); + } + return TINYEXR_ERROR_INVALID_HEADER; + } + + size_t num_tiles = offsets.size(); // = # of blocks + + exr_image->tiles = static_cast( + calloc(static_cast(num_tiles), sizeof(EXRTile))); + + int err_code = TINYEXR_SUCCESS; + +#if (__cplusplus > 199711L) && (TINYEXR_USE_THREAD > 0) + + std::vector workers; + std::atomic tile_count(0); + + int num_threads = std::max(1, int(std::thread::hardware_concurrency())); + if (num_threads > int(num_tiles)) { + num_threads = int(num_tiles); + } + + for (int t = 0; t < num_threads; t++) { + workers.emplace_back(std::thread([&]() { + size_t tile_idx = 0; + while ((tile_idx = tile_count++) < num_tiles) { + +#else + for (size_t tile_idx = 0; tile_idx < num_tiles; tile_idx++) { +#endif + // Allocate memory for each tile. + exr_image->tiles[tile_idx].images = tinyexr::AllocateImage( + num_channels, exr_header->channels, + exr_header->requested_pixel_types, exr_header->tile_size_x, + exr_header->tile_size_y); + + // 16 byte: tile coordinates + // 4 byte : data size + // ~ : data(uncompressed or compressed) + if (offsets[tile_idx] + sizeof(int) * 5 > size) { + // TODO(LTE): atomic + if (err) { + (*err) += "Insufficient data size.\n"; + } + err_code = TINYEXR_ERROR_INVALID_DATA; + break; + } + + size_t data_size = + size_t(size - (offsets[tile_idx] + sizeof(int) * 5)); + const unsigned char *data_ptr = + reinterpret_cast(head + offsets[tile_idx]); + + int tile_coordinates[4]; + memcpy(tile_coordinates, data_ptr, sizeof(int) * 4); + tinyexr::swap4( + reinterpret_cast(&tile_coordinates[0])); + tinyexr::swap4( + reinterpret_cast(&tile_coordinates[1])); + tinyexr::swap4( + reinterpret_cast(&tile_coordinates[2])); + tinyexr::swap4( + reinterpret_cast(&tile_coordinates[3])); + + // @todo{ LoD } + if (tile_coordinates[2] != 0) { + err_code = TINYEXR_ERROR_UNSUPPORTED_FEATURE; + break; + } + if (tile_coordinates[3] != 0) { + err_code = TINYEXR_ERROR_UNSUPPORTED_FEATURE; + break; + } + + int data_len; + memcpy(&data_len, data_ptr + 16, + sizeof(int)); // 16 = sizeof(tile_coordinates) + tinyexr::swap4(reinterpret_cast(&data_len)); + + if (data_len < 4 || size_t(data_len) > data_size) { + // TODO(LTE): atomic + if (err) { + (*err) += "Insufficient data length.\n"; + } + err_code = TINYEXR_ERROR_INVALID_DATA; + break; + } + + // Move to data addr: 20 = 16 + 4; + data_ptr += 20; + + bool ret = tinyexr::DecodeTiledPixelData( + exr_image->tiles[tile_idx].images, + &(exr_image->tiles[tile_idx].width), + &(exr_image->tiles[tile_idx].height), + exr_header->requested_pixel_types, data_ptr, + static_cast(data_len), exr_header->compression_type, + exr_header->line_order, data_width, data_height, + tile_coordinates[0], tile_coordinates[1], exr_header->tile_size_x, + exr_header->tile_size_y, static_cast(pixel_data_size), + static_cast(exr_header->num_custom_attributes), + exr_header->custom_attributes, + static_cast(exr_header->num_channels), + exr_header->channels, channel_offset_list); + + if (!ret) { + // TODO(LTE): atomic + if (err) { + (*err) += "Failed to decode tile data.\n"; + } + err_code = TINYEXR_ERROR_INVALID_DATA; + } + + exr_image->tiles[tile_idx].offset_x = tile_coordinates[0]; + exr_image->tiles[tile_idx].offset_y = tile_coordinates[1]; + exr_image->tiles[tile_idx].level_x = tile_coordinates[2]; + exr_image->tiles[tile_idx].level_y = tile_coordinates[3]; + +#if (__cplusplus > 199711L) && (TINYEXR_USE_THREAD > 0) + } + })); + } // num_thread loop + + for (auto &t : workers) { + t.join(); + } + +#else + } +#endif + + if (err_code != TINYEXR_SUCCESS) { + return err_code; + } + + exr_image->num_tiles = static_cast(num_tiles); + } else { // scanline format + + // Don't allow too large image(256GB * pixel_data_size or more). Workaround + // for #104. + size_t total_data_len = + size_t(data_width) * size_t(data_height) * size_t(num_channels); + const bool total_data_len_overflown = + sizeof(void *) == 8 ? (total_data_len >= 0x4000000000) : false; + if ((total_data_len == 0) || total_data_len_overflown) { + if (err) { + std::stringstream ss; + ss << "Image data size is zero or too large: width = " << data_width + << ", height = " << data_height << ", channels = " << num_channels + << std::endl; + (*err) += ss.str(); + } + return TINYEXR_ERROR_INVALID_DATA; + } + + exr_image->images = tinyexr::AllocateImage( + num_channels, exr_header->channels, exr_header->requested_pixel_types, + data_width, data_height); + +#if (__cplusplus > 199711L) && (TINYEXR_USE_THREAD > 0) + std::vector workers; + std::atomic y_count(0); + + int num_threads = std::max(1, int(std::thread::hardware_concurrency())); + if (num_threads > int(num_blocks)) { + num_threads = int(num_blocks); + } + + for (int t = 0; t < num_threads; t++) { + workers.emplace_back(std::thread([&]() { + int y = 0; + while ((y = y_count++) < int(num_blocks)) { + +#else + +#if TINYEXR_USE_OPENMP +#pragma omp parallel for +#endif + for (int y = 0; y < static_cast(num_blocks); y++) { + +#endif + size_t y_idx = static_cast(y); + + if (offsets[y_idx] + sizeof(int) * 2 > size) { + invalid_data = true; + } else { + // 4 byte: scan line + // 4 byte: data size + // ~ : pixel data(uncompressed or compressed) + size_t data_size = + size_t(size - (offsets[y_idx] + sizeof(int) * 2)); + const unsigned char *data_ptr = + reinterpret_cast(head + offsets[y_idx]); + + int line_no; + memcpy(&line_no, data_ptr, sizeof(int)); + int data_len; + memcpy(&data_len, data_ptr + 4, sizeof(int)); + tinyexr::swap4(reinterpret_cast(&line_no)); + tinyexr::swap4(reinterpret_cast(&data_len)); + + if (size_t(data_len) > data_size) { + invalid_data = true; + + } else if ((line_no > (2 << 20)) || (line_no < -(2 << 20))) { + // Too large value. Assume this is invalid + // 2**20 = 1048576 = heuristic value. + invalid_data = true; + } else if (data_len == 0) { + // TODO(syoyo): May be ok to raise the threshold for example + // `data_len < 4` + invalid_data = true; + } else { + // line_no may be negative. + int end_line_no = (std::min)(line_no + num_scanline_blocks, + (exr_header->data_window[3] + 1)); + + int num_lines = end_line_no - line_no; + + if (num_lines <= 0) { + invalid_data = true; + } else { + // Move to data addr: 8 = 4 + 4; + data_ptr += 8; + + // Adjust line_no with data_window.bmin.y + + // overflow check + tinyexr_int64 lno = + static_cast(line_no) - + static_cast(exr_header->data_window[1]); + if (lno > std::numeric_limits::max()) { + line_no = -1; // invalid + } else if (lno < -std::numeric_limits::max()) { + line_no = -1; // invalid + } else { + line_no -= exr_header->data_window[1]; + } + + if (line_no < 0) { + invalid_data = true; + } else { + if (!tinyexr::DecodePixelData( + exr_image->images, exr_header->requested_pixel_types, + data_ptr, static_cast(data_len), + exr_header->compression_type, exr_header->line_order, + data_width, data_height, data_width, y, line_no, + num_lines, static_cast(pixel_data_size), + static_cast( + exr_header->num_custom_attributes), + exr_header->custom_attributes, + static_cast(exr_header->num_channels), + exr_header->channels, channel_offset_list)) { + invalid_data = true; + } + } + } + } + } + +#if (__cplusplus > 199711L) && (TINYEXR_USE_THREAD > 0) + } + })); + } + + for (auto &t : workers) { + t.join(); + } +#else + } // omp parallel +#endif + } + + if (invalid_data) { + if (err) { + std::stringstream ss; + (*err) += "Invalid data found when decoding pixels.\n"; + } + return TINYEXR_ERROR_INVALID_DATA; + } + + // Overwrite `pixel_type` with `requested_pixel_type`. + { + for (int c = 0; c < exr_header->num_channels; c++) { + exr_header->pixel_types[c] = exr_header->requested_pixel_types[c]; + } + } + + { + exr_image->num_channels = num_channels; + + exr_image->width = data_width; + exr_image->height = data_height; + } + + return TINYEXR_SUCCESS; +} + +static bool ReconstructLineOffsets( + std::vector *offsets, size_t n, + const unsigned char *head, const unsigned char *marker, const size_t size) { + TEXR_ASSERT(head < marker); + TEXR_ASSERT(offsets->size() == n); + + for (size_t i = 0; i < n; i++) { + size_t offset = static_cast(marker - head); + // Offset should not exceed whole EXR file/data size. + if ((offset + sizeof(tinyexr::tinyexr_uint64)) >= size) { + return false; + } + + int y; + unsigned int data_len; + + memcpy(&y, marker, sizeof(int)); + memcpy(&data_len, marker + 4, sizeof(unsigned int)); + + if (data_len >= size) { + return false; + } + + tinyexr::swap4(reinterpret_cast(&y)); + tinyexr::swap4(reinterpret_cast(&data_len)); + + (*offsets)[i] = offset; + + marker += data_len + 8; // 8 = 4 bytes(y) + 4 bytes(data_len) + } + + return true; +} + +static int DecodeEXRImage(EXRImage *exr_image, const EXRHeader *exr_header, + const unsigned char *head, + const unsigned char *marker, const size_t size, + const char **err) { + if (exr_image == NULL || exr_header == NULL || head == NULL || + marker == NULL || (size <= tinyexr::kEXRVersionSize)) { + tinyexr::SetErrorMessage("Invalid argument for DecodeEXRImage().", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + int num_scanline_blocks = 1; + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZIP) { + num_scanline_blocks = 16; + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { + num_scanline_blocks = 32; + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { + num_scanline_blocks = 16; + } + + int data_width = exr_header->data_window[2] - exr_header->data_window[0]; + if (data_width >= std::numeric_limits::max()) { + // Issue 63 + tinyexr::SetErrorMessage("Invalid data width value", err); + return TINYEXR_ERROR_INVALID_DATA; + } + data_width++; + + int data_height = exr_header->data_window[3] - exr_header->data_window[1]; + if (data_height >= std::numeric_limits::max()) { + tinyexr::SetErrorMessage("Invalid data height value", err); + return TINYEXR_ERROR_INVALID_DATA; + } + data_height++; + + if ((data_width < 0) || (data_height < 0)) { + tinyexr::SetErrorMessage("data width or data height is negative.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + // Do not allow too large data_width and data_height. header invalid? + { + const int threshold = 1024 * 8192; // heuristics + if (data_width > threshold) { + tinyexr::SetErrorMessage("data width too large.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + if (data_height > threshold) { + tinyexr::SetErrorMessage("data height too large.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + } + + // Read offset tables. + size_t num_blocks = 0; + + if (exr_header->chunk_count > 0) { + // Use `chunkCount` attribute. + num_blocks = static_cast(exr_header->chunk_count); + } else if (exr_header->tiled) { + // @todo { LoD } + size_t num_x_tiles = static_cast(data_width) / + static_cast(exr_header->tile_size_x); + if (num_x_tiles * static_cast(exr_header->tile_size_x) < + static_cast(data_width)) { + num_x_tiles++; + } + size_t num_y_tiles = static_cast(data_height) / + static_cast(exr_header->tile_size_y); + if (num_y_tiles * static_cast(exr_header->tile_size_y) < + static_cast(data_height)) { + num_y_tiles++; + } + + num_blocks = num_x_tiles * num_y_tiles; + } else { + num_blocks = static_cast(data_height) / + static_cast(num_scanline_blocks); + if (num_blocks * static_cast(num_scanline_blocks) < + static_cast(data_height)) { + num_blocks++; + } + } + + std::vector offsets(num_blocks); + + for (size_t y = 0; y < num_blocks; y++) { + tinyexr::tinyexr_uint64 offset; + // Issue #81 + if ((marker + sizeof(tinyexr_uint64)) >= (head + size)) { + tinyexr::SetErrorMessage("Insufficient data size in offset table.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + memcpy(&offset, marker, sizeof(tinyexr::tinyexr_uint64)); + tinyexr::swap8(&offset); + if (offset >= size) { + tinyexr::SetErrorMessage("Invalid offset value in DecodeEXRImage.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + marker += sizeof(tinyexr::tinyexr_uint64); // = 8 + offsets[y] = offset; + } + + // If line offsets are invalid, we try to reconstruct it. + // See OpenEXR/IlmImf/ImfScanLineInputFile.cpp::readLineOffsets() for details. + for (size_t y = 0; y < num_blocks; y++) { + if (offsets[y] <= 0) { + // TODO(syoyo) Report as warning? + // if (err) { + // stringstream ss; + // ss << "Incomplete lineOffsets." << std::endl; + // (*err) += ss.str(); + //} + bool ret = + ReconstructLineOffsets(&offsets, num_blocks, head, marker, size); + if (ret) { + // OK + break; + } else { + tinyexr::SetErrorMessage( + "Cannot reconstruct lineOffset table in DecodeEXRImage.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + } + } + + { + std::string e; + int ret = DecodeChunk(exr_image, exr_header, offsets, head, size, &e); + + if (ret != TINYEXR_SUCCESS) { + if (!e.empty()) { + tinyexr::SetErrorMessage(e, err); + } + +#if 1 + FreeEXRImage(exr_image); +#else + // release memory(if exists) + if ((exr_header->num_channels > 0) && exr_image && exr_image->images) { + for (size_t c = 0; c < size_t(exr_header->num_channels); c++) { + if (exr_image->images[c]) { + free(exr_image->images[c]); + exr_image->images[c] = NULL; + } + } + free(exr_image->images); + exr_image->images = NULL; + } +#endif + } + + return ret; + } +} + +static void GetLayers(const EXRHeader& exr_header, std::vector& layer_names) { + // Naive implementation + // Group channels by layers + // go over all channel names, split by periods + // collect unique names + layer_names.clear(); + for (int c = 0; c < exr_header.num_channels; c++) { + std::string full_name(exr_header.channels[c].name); + const size_t pos = full_name.find_last_of('.'); + if (pos != std::string::npos && pos != 0 && pos + 1 < full_name.size()) { + full_name.erase(pos); + if (std::find(layer_names.begin(), layer_names.end(), full_name) == layer_names.end()) + layer_names.push_back(full_name); + } + } +} + +struct LayerChannel { + explicit LayerChannel (size_t i, std::string n) + : index(i) + , name(n) + {} + size_t index; + std::string name; +}; + +static void ChannelsInLayer(const EXRHeader& exr_header, const std::string layer_name, std::vector& channels) { + channels.clear(); + for (int c = 0; c < exr_header.num_channels; c++) { + std::string ch_name(exr_header.channels[c].name); + if (layer_name.empty()) { + const size_t pos = ch_name.find_last_of('.'); + if (pos != std::string::npos && pos < ch_name.size()) { + ch_name = ch_name.substr(pos + 1); + } + } else { + const size_t pos = ch_name.find(layer_name + '.'); + if (pos == std::string::npos) + continue; + if (pos == 0) { + ch_name = ch_name.substr(layer_name.size() + 1); + } + } + LayerChannel ch(size_t(c), ch_name); + channels.push_back(ch); + } +} + +} // namespace tinyexr + +int EXRLayers(const char *filename, const char **layer_names[], int *num_layers, const char **err) { + EXRVersion exr_version; + EXRHeader exr_header; + InitEXRHeader(&exr_header); + + { + int ret = ParseEXRVersionFromFile(&exr_version, filename); + if (ret != TINYEXR_SUCCESS) { + tinyexr::SetErrorMessage("Invalid EXR header.", err); + return ret; + } + + if (exr_version.multipart || exr_version.non_image) { + tinyexr::SetErrorMessage( + "Loading multipart or DeepImage is not supported in LoadEXR() API", + err); + return TINYEXR_ERROR_INVALID_DATA; // @fixme. + } + } + + int ret = ParseEXRHeaderFromFile(&exr_header, &exr_version, filename, err); + if (ret != TINYEXR_SUCCESS) { + FreeEXRHeader(&exr_header); + return ret; + } + + std::vector layer_vec; + tinyexr::GetLayers(exr_header, layer_vec); + + (*num_layers) = int(layer_vec.size()); + (*layer_names) = static_cast( + malloc(sizeof(const char *) * static_cast(layer_vec.size()))); + for (size_t c = 0; c < static_cast(layer_vec.size()); c++) { +#ifdef _MSC_VER + (*layer_names)[c] = _strdup(layer_vec[c].c_str()); +#else + (*layer_names)[c] = strdup(layer_vec[c].c_str()); +#endif + } + + FreeEXRHeader(&exr_header); + return TINYEXR_SUCCESS; +} + +int LoadEXR(float **out_rgba, int *width, int *height, const char *filename, + const char **err) { + return LoadEXRWithLayer(out_rgba, width, height, filename, /* layername */NULL, err); +} + +int LoadEXRWithLayer(float **out_rgba, int *width, int *height, const char *filename, const char *layername, + const char **err) { + if (out_rgba == NULL) { + tinyexr::SetErrorMessage("Invalid argument for LoadEXR()", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + EXRVersion exr_version; + EXRImage exr_image; + EXRHeader exr_header; + InitEXRHeader(&exr_header); + InitEXRImage(&exr_image); + + { + int ret = ParseEXRVersionFromFile(&exr_version, filename); + if (ret != TINYEXR_SUCCESS) { + std::stringstream ss; + ss << "Failed to open EXR file or read version info from EXR file. code(" << ret << ")"; + tinyexr::SetErrorMessage(ss.str(), err); + return ret; + } + + if (exr_version.multipart || exr_version.non_image) { + tinyexr::SetErrorMessage( + "Loading multipart or DeepImage is not supported in LoadEXR() API", + err); + return TINYEXR_ERROR_INVALID_DATA; // @fixme. + } + } + + { + int ret = ParseEXRHeaderFromFile(&exr_header, &exr_version, filename, err); + if (ret != TINYEXR_SUCCESS) { + FreeEXRHeader(&exr_header); + return ret; + } + } + + // Read HALF channel as FLOAT. + for (int i = 0; i < exr_header.num_channels; i++) { + if (exr_header.pixel_types[i] == TINYEXR_PIXELTYPE_HALF) { + exr_header.requested_pixel_types[i] = TINYEXR_PIXELTYPE_FLOAT; + } + } + + // TODO: Probably limit loading to layers (channels) selected by layer index + { + int ret = LoadEXRImageFromFile(&exr_image, &exr_header, filename, err); + if (ret != TINYEXR_SUCCESS) { + FreeEXRHeader(&exr_header); + return ret; + } + } + + // RGBA + int idxR = -1; + int idxG = -1; + int idxB = -1; + int idxA = -1; + + std::vector layer_names; + tinyexr::GetLayers(exr_header, layer_names); + + std::vector channels; + tinyexr::ChannelsInLayer(exr_header, layername == NULL ? "" : std::string(layername), channels); + + if (channels.size() < 1) { + tinyexr::SetErrorMessage("Layer Not Found", err); + FreeEXRHeader(&exr_header); + FreeEXRImage(&exr_image); + return TINYEXR_ERROR_LAYER_NOT_FOUND; + } + + size_t ch_count = channels.size() < 4 ? channels.size() : 4; + for (size_t c = 0; c < ch_count; c++) { + const tinyexr::LayerChannel &ch = channels[c]; + + if (ch.name == "R") { + idxR = int(ch.index); + } + else if (ch.name == "G") { + idxG = int(ch.index); + } + else if (ch.name == "B") { + idxB = int(ch.index); + } + else if (ch.name == "A") { + idxA = int(ch.index); + } + } + + if (channels.size() == 1) { + int chIdx = int(channels.front().index); + // Grayscale channel only. + + (*out_rgba) = reinterpret_cast( + malloc(4 * sizeof(float) * static_cast(exr_image.width) * + static_cast(exr_image.height))); + + if (exr_header.tiled) { + for (int it = 0; it < exr_image.num_tiles; it++) { + for (int j = 0; j < exr_header.tile_size_y; j++) { + for (int i = 0; i < exr_header.tile_size_x; i++) { + const int ii = + exr_image.tiles[it].offset_x * exr_header.tile_size_x + i; + const int jj = + exr_image.tiles[it].offset_y * exr_header.tile_size_y + j; + const int idx = ii + jj * exr_image.width; + + // out of region check. + if (ii >= exr_image.width) { + continue; + } + if (jj >= exr_image.height) { + continue; + } + const int srcIdx = i + j * exr_header.tile_size_x; + unsigned char **src = exr_image.tiles[it].images; + (*out_rgba)[4 * idx + 0] = + reinterpret_cast(src)[chIdx][srcIdx]; + (*out_rgba)[4 * idx + 1] = + reinterpret_cast(src)[chIdx][srcIdx]; + (*out_rgba)[4 * idx + 2] = + reinterpret_cast(src)[chIdx][srcIdx]; + (*out_rgba)[4 * idx + 3] = + reinterpret_cast(src)[chIdx][srcIdx]; + } + } + } + } else { + for (int i = 0; i < exr_image.width * exr_image.height; i++) { + const float val = reinterpret_cast(exr_image.images)[chIdx][i]; + (*out_rgba)[4 * i + 0] = val; + (*out_rgba)[4 * i + 1] = val; + (*out_rgba)[4 * i + 2] = val; + (*out_rgba)[4 * i + 3] = val; + } + } + } else { + // Assume RGB(A) + + if (idxR == -1) { + tinyexr::SetErrorMessage("R channel not found", err); + + FreeEXRHeader(&exr_header); + FreeEXRImage(&exr_image); + return TINYEXR_ERROR_INVALID_DATA; + } + + if (idxG == -1) { + tinyexr::SetErrorMessage("G channel not found", err); + FreeEXRHeader(&exr_header); + FreeEXRImage(&exr_image); + return TINYEXR_ERROR_INVALID_DATA; + } + + if (idxB == -1) { + tinyexr::SetErrorMessage("B channel not found", err); + FreeEXRHeader(&exr_header); + FreeEXRImage(&exr_image); + return TINYEXR_ERROR_INVALID_DATA; + } + + (*out_rgba) = reinterpret_cast( + malloc(4 * sizeof(float) * static_cast(exr_image.width) * + static_cast(exr_image.height))); + if (exr_header.tiled) { + for (int it = 0; it < exr_image.num_tiles; it++) { + for (int j = 0; j < exr_header.tile_size_y; j++) { + for (int i = 0; i < exr_header.tile_size_x; i++) { + const int ii = + exr_image.tiles[it].offset_x * exr_header.tile_size_x + i; + const int jj = + exr_image.tiles[it].offset_y * exr_header.tile_size_y + j; + const int idx = ii + jj * exr_image.width; + + // out of region check. + if (ii >= exr_image.width) { + continue; + } + if (jj >= exr_image.height) { + continue; + } + const int srcIdx = i + j * exr_header.tile_size_x; + unsigned char **src = exr_image.tiles[it].images; + (*out_rgba)[4 * idx + 0] = + reinterpret_cast(src)[idxR][srcIdx]; + (*out_rgba)[4 * idx + 1] = + reinterpret_cast(src)[idxG][srcIdx]; + (*out_rgba)[4 * idx + 2] = + reinterpret_cast(src)[idxB][srcIdx]; + if (idxA != -1) { + (*out_rgba)[4 * idx + 3] = + reinterpret_cast(src)[idxA][srcIdx]; + } else { + (*out_rgba)[4 * idx + 3] = 1.0; + } + } + } + } + } else { + for (int i = 0; i < exr_image.width * exr_image.height; i++) { + (*out_rgba)[4 * i + 0] = + reinterpret_cast(exr_image.images)[idxR][i]; + (*out_rgba)[4 * i + 1] = + reinterpret_cast(exr_image.images)[idxG][i]; + (*out_rgba)[4 * i + 2] = + reinterpret_cast(exr_image.images)[idxB][i]; + if (idxA != -1) { + (*out_rgba)[4 * i + 3] = + reinterpret_cast(exr_image.images)[idxA][i]; + } else { + (*out_rgba)[4 * i + 3] = 1.0; + } + } + } + } + + (*width) = exr_image.width; + (*height) = exr_image.height; + + FreeEXRHeader(&exr_header); + FreeEXRImage(&exr_image); + + return TINYEXR_SUCCESS; +} + +int IsEXR(const char *filename) { + EXRVersion exr_version; + + int ret = ParseEXRVersionFromFile(&exr_version, filename); + if (ret != TINYEXR_SUCCESS) { + return ret; + } + + return TINYEXR_SUCCESS; +} + +int ParseEXRHeaderFromMemory(EXRHeader *exr_header, const EXRVersion *version, + const unsigned char *memory, size_t size, + const char **err) { + if (memory == NULL || exr_header == NULL) { + tinyexr::SetErrorMessage( + "Invalid argument. `memory` or `exr_header` argument is null in " + "ParseEXRHeaderFromMemory()", + err); + + // Invalid argument + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + if (size < tinyexr::kEXRVersionSize) { + tinyexr::SetErrorMessage("Insufficient header/data size.\n", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + const unsigned char *marker = memory + tinyexr::kEXRVersionSize; + size_t marker_size = size - tinyexr::kEXRVersionSize; + + tinyexr::HeaderInfo info; + info.clear(); + + std::string err_str; + int ret = ParseEXRHeader(&info, NULL, version, &err_str, marker, marker_size); + + if (ret != TINYEXR_SUCCESS) { + if (err && !err_str.empty()) { + tinyexr::SetErrorMessage(err_str, err); + } + } + + ConvertHeader(exr_header, info); + + // transfoer `tiled` from version. + exr_header->tiled = version->tiled; + + return ret; +} + +int LoadEXRFromMemory(float **out_rgba, int *width, int *height, + const unsigned char *memory, size_t size, + const char **err) { + if (out_rgba == NULL || memory == NULL) { + tinyexr::SetErrorMessage("Invalid argument for LoadEXRFromMemory", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + EXRVersion exr_version; + EXRImage exr_image; + EXRHeader exr_header; + + InitEXRHeader(&exr_header); + + int ret = ParseEXRVersionFromMemory(&exr_version, memory, size); + if (ret != TINYEXR_SUCCESS) { + std::stringstream ss; + ss << "Failed to parse EXR version. code(" << ret << ")"; + tinyexr::SetErrorMessage(ss.str(), err); + return ret; + } + + ret = ParseEXRHeaderFromMemory(&exr_header, &exr_version, memory, size, err); + if (ret != TINYEXR_SUCCESS) { + return ret; + } + + // Read HALF channel as FLOAT. + for (int i = 0; i < exr_header.num_channels; i++) { + if (exr_header.pixel_types[i] == TINYEXR_PIXELTYPE_HALF) { + exr_header.requested_pixel_types[i] = TINYEXR_PIXELTYPE_FLOAT; + } + } + + InitEXRImage(&exr_image); + ret = LoadEXRImageFromMemory(&exr_image, &exr_header, memory, size, err); + if (ret != TINYEXR_SUCCESS) { + return ret; + } + + // RGBA + int idxR = -1; + int idxG = -1; + int idxB = -1; + int idxA = -1; + for (int c = 0; c < exr_header.num_channels; c++) { + if (strcmp(exr_header.channels[c].name, "R") == 0) { + idxR = c; + } else if (strcmp(exr_header.channels[c].name, "G") == 0) { + idxG = c; + } else if (strcmp(exr_header.channels[c].name, "B") == 0) { + idxB = c; + } else if (strcmp(exr_header.channels[c].name, "A") == 0) { + idxA = c; + } + } + + // TODO(syoyo): Refactor removing same code as used in LoadEXR(). + if (exr_header.num_channels == 1) { + // Grayscale channel only. + + (*out_rgba) = reinterpret_cast( + malloc(4 * sizeof(float) * static_cast(exr_image.width) * + static_cast(exr_image.height))); + + if (exr_header.tiled) { + for (int it = 0; it < exr_image.num_tiles; it++) { + for (int j = 0; j < exr_header.tile_size_y; j++) { + for (int i = 0; i < exr_header.tile_size_x; i++) { + const int ii = + exr_image.tiles[it].offset_x * exr_header.tile_size_x + i; + const int jj = + exr_image.tiles[it].offset_y * exr_header.tile_size_y + j; + const int idx = ii + jj * exr_image.width; + + // out of region check. + if (ii >= exr_image.width) { + continue; + } + if (jj >= exr_image.height) { + continue; + } + const int srcIdx = i + j * exr_header.tile_size_x; + unsigned char **src = exr_image.tiles[it].images; + (*out_rgba)[4 * idx + 0] = + reinterpret_cast(src)[0][srcIdx]; + (*out_rgba)[4 * idx + 1] = + reinterpret_cast(src)[0][srcIdx]; + (*out_rgba)[4 * idx + 2] = + reinterpret_cast(src)[0][srcIdx]; + (*out_rgba)[4 * idx + 3] = + reinterpret_cast(src)[0][srcIdx]; + } + } + } + } else { + for (int i = 0; i < exr_image.width * exr_image.height; i++) { + const float val = reinterpret_cast(exr_image.images)[0][i]; + (*out_rgba)[4 * i + 0] = val; + (*out_rgba)[4 * i + 1] = val; + (*out_rgba)[4 * i + 2] = val; + (*out_rgba)[4 * i + 3] = val; + } + } + + } else { + // TODO(syoyo): Support non RGBA image. + + if (idxR == -1) { + tinyexr::SetErrorMessage("R channel not found", err); + + // @todo { free exr_image } + return TINYEXR_ERROR_INVALID_DATA; + } + + if (idxG == -1) { + tinyexr::SetErrorMessage("G channel not found", err); + // @todo { free exr_image } + return TINYEXR_ERROR_INVALID_DATA; + } + + if (idxB == -1) { + tinyexr::SetErrorMessage("B channel not found", err); + // @todo { free exr_image } + return TINYEXR_ERROR_INVALID_DATA; + } + + (*out_rgba) = reinterpret_cast( + malloc(4 * sizeof(float) * static_cast(exr_image.width) * + static_cast(exr_image.height))); + + if (exr_header.tiled) { + for (int it = 0; it < exr_image.num_tiles; it++) { + for (int j = 0; j < exr_header.tile_size_y; j++) + for (int i = 0; i < exr_header.tile_size_x; i++) { + const int ii = + exr_image.tiles[it].offset_x * exr_header.tile_size_x + i; + const int jj = + exr_image.tiles[it].offset_y * exr_header.tile_size_y + j; + const int idx = ii + jj * exr_image.width; + + // out of region check. + if (ii >= exr_image.width) { + continue; + } + if (jj >= exr_image.height) { + continue; + } + const int srcIdx = i + j * exr_header.tile_size_x; + unsigned char **src = exr_image.tiles[it].images; + (*out_rgba)[4 * idx + 0] = + reinterpret_cast(src)[idxR][srcIdx]; + (*out_rgba)[4 * idx + 1] = + reinterpret_cast(src)[idxG][srcIdx]; + (*out_rgba)[4 * idx + 2] = + reinterpret_cast(src)[idxB][srcIdx]; + if (idxA != -1) { + (*out_rgba)[4 * idx + 3] = + reinterpret_cast(src)[idxA][srcIdx]; + } else { + (*out_rgba)[4 * idx + 3] = 1.0; + } + } + } + } else { + for (int i = 0; i < exr_image.width * exr_image.height; i++) { + (*out_rgba)[4 * i + 0] = + reinterpret_cast(exr_image.images)[idxR][i]; + (*out_rgba)[4 * i + 1] = + reinterpret_cast(exr_image.images)[idxG][i]; + (*out_rgba)[4 * i + 2] = + reinterpret_cast(exr_image.images)[idxB][i]; + if (idxA != -1) { + (*out_rgba)[4 * i + 3] = + reinterpret_cast(exr_image.images)[idxA][i]; + } else { + (*out_rgba)[4 * i + 3] = 1.0; + } + } + } + } + + (*width) = exr_image.width; + (*height) = exr_image.height; + + FreeEXRHeader(&exr_header); + FreeEXRImage(&exr_image); + + return TINYEXR_SUCCESS; +} + +int LoadEXRImageFromFile(EXRImage *exr_image, const EXRHeader *exr_header, + const char *filename, const char **err) { + if (exr_image == NULL) { + tinyexr::SetErrorMessage("Invalid argument for LoadEXRImageFromFile", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#ifdef _WIN32 + FILE *fp = NULL; + fopen_s(&fp, filename, "rb"); +#else + FILE *fp = fopen(filename, "rb"); +#endif + if (!fp) { + tinyexr::SetErrorMessage("Cannot read file " + std::string(filename), err); + return TINYEXR_ERROR_CANT_OPEN_FILE; + } + + size_t filesize; + // Compute size + fseek(fp, 0, SEEK_END); + filesize = static_cast(ftell(fp)); + fseek(fp, 0, SEEK_SET); + + if (filesize < 16) { + fclose(fp); + tinyexr::SetErrorMessage("File size too short " + std::string(filename), + err); + return TINYEXR_ERROR_INVALID_FILE; + } + + std::vector buf(filesize); // @todo { use mmap } + { + size_t ret; + ret = fread(&buf[0], 1, filesize, fp); + TEXR_ASSERT(ret == filesize); + fclose(fp); + (void)ret; + } + + return LoadEXRImageFromMemory(exr_image, exr_header, &buf.at(0), filesize, + err); +} + +int LoadEXRImageFromMemory(EXRImage *exr_image, const EXRHeader *exr_header, + const unsigned char *memory, const size_t size, + const char **err) { + if (exr_image == NULL || memory == NULL || + (size < tinyexr::kEXRVersionSize)) { + tinyexr::SetErrorMessage("Invalid argument for LoadEXRImageFromMemory", + err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + if (exr_header->header_len == 0) { + tinyexr::SetErrorMessage("EXRHeader variable is not initialized.", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + const unsigned char *head = memory; + const unsigned char *marker = reinterpret_cast( + memory + exr_header->header_len + + 8); // +8 for magic number + version header. + return tinyexr::DecodeEXRImage(exr_image, exr_header, head, marker, size, + err); +} + +size_t SaveEXRImageToMemory(const EXRImage *exr_image, + const EXRHeader *exr_header, + unsigned char **memory_out, const char **err) { + if (exr_image == NULL || memory_out == NULL || + exr_header->compression_type < 0) { + tinyexr::SetErrorMessage("Invalid argument for SaveEXRImageToMemory", err); + return 0; + } + +#if !TINYEXR_USE_PIZ + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { + tinyexr::SetErrorMessage("PIZ compression is not supported in this build", + err); + return 0; + } +#endif + +#if !TINYEXR_USE_ZFP + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { + tinyexr::SetErrorMessage("ZFP compression is not supported in this build", + err); + return 0; + } +#endif + +#if TINYEXR_USE_ZFP + for (size_t i = 0; i < static_cast(exr_header->num_channels); i++) { + if (exr_header->requested_pixel_types[i] != TINYEXR_PIXELTYPE_FLOAT) { + tinyexr::SetErrorMessage("Pixel type must be FLOAT for ZFP compression", + err); + return 0; + } + } +#endif + + std::vector memory; + + // Header + { + const char header[] = {0x76, 0x2f, 0x31, 0x01}; + memory.insert(memory.end(), header, header + 4); + } + + // Version, scanline. + { + char marker[] = {2, 0, 0, 0}; + /* @todo + if (exr_header->tiled) { + marker[1] |= 0x2; + } + if (exr_header->long_name) { + marker[1] |= 0x4; + } + if (exr_header->non_image) { + marker[1] |= 0x8; + } + if (exr_header->multipart) { + marker[1] |= 0x10; + } + */ + memory.insert(memory.end(), marker, marker + 4); + } + + int num_scanlines = 1; + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZIP) { + num_scanlines = 16; + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { + num_scanlines = 32; + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { + num_scanlines = 16; + } + + // Write attributes. + std::vector channels; + { + std::vector data; + + for (int c = 0; c < exr_header->num_channels; c++) { + tinyexr::ChannelInfo info; + info.p_linear = 0; + info.pixel_type = exr_header->requested_pixel_types[c]; + info.x_sampling = 1; + info.y_sampling = 1; + info.name = std::string(exr_header->channels[c].name); + channels.push_back(info); + } + + tinyexr::WriteChannelInfo(data, channels); + + tinyexr::WriteAttributeToMemory(&memory, "channels", "chlist", &data.at(0), + static_cast(data.size())); + } + + { + int comp = exr_header->compression_type; + tinyexr::swap4(reinterpret_cast(&comp)); + tinyexr::WriteAttributeToMemory( + &memory, "compression", "compression", + reinterpret_cast(&comp), 1); + } + + { + int data[4] = {0, 0, exr_image->width - 1, exr_image->height - 1}; + tinyexr::swap4(reinterpret_cast(&data[0])); + tinyexr::swap4(reinterpret_cast(&data[1])); + tinyexr::swap4(reinterpret_cast(&data[2])); + tinyexr::swap4(reinterpret_cast(&data[3])); + tinyexr::WriteAttributeToMemory( + &memory, "dataWindow", "box2i", + reinterpret_cast(data), sizeof(int) * 4); + tinyexr::WriteAttributeToMemory( + &memory, "displayWindow", "box2i", + reinterpret_cast(data), sizeof(int) * 4); + } + + { + unsigned char line_order = 0; // @fixme { read line_order from EXRHeader } + tinyexr::WriteAttributeToMemory(&memory, "lineOrder", "lineOrder", + &line_order, 1); + } + + { + float aspectRatio = 1.0f; + tinyexr::swap4(reinterpret_cast(&aspectRatio)); + tinyexr::WriteAttributeToMemory( + &memory, "pixelAspectRatio", "float", + reinterpret_cast(&aspectRatio), sizeof(float)); + } + + { + float center[2] = {0.0f, 0.0f}; + tinyexr::swap4(reinterpret_cast(¢er[0])); + tinyexr::swap4(reinterpret_cast(¢er[1])); + tinyexr::WriteAttributeToMemory( + &memory, "screenWindowCenter", "v2f", + reinterpret_cast(center), 2 * sizeof(float)); + } + + { + float w = static_cast(exr_image->width); + tinyexr::swap4(reinterpret_cast(&w)); + tinyexr::WriteAttributeToMemory(&memory, "screenWindowWidth", "float", + reinterpret_cast(&w), + sizeof(float)); + } + + // Custom attributes + if (exr_header->num_custom_attributes > 0) { + for (int i = 0; i < exr_header->num_custom_attributes; i++) { + tinyexr::WriteAttributeToMemory( + &memory, exr_header->custom_attributes[i].name, + exr_header->custom_attributes[i].type, + reinterpret_cast( + exr_header->custom_attributes[i].value), + exr_header->custom_attributes[i].size); + } + } + + { // end of header + unsigned char e = 0; + memory.push_back(e); + } + + int num_blocks = exr_image->height / num_scanlines; + if (num_blocks * num_scanlines < exr_image->height) { + num_blocks++; + } + + std::vector offsets(static_cast(num_blocks)); + + size_t headerSize = memory.size(); + tinyexr::tinyexr_uint64 offset = + headerSize + + static_cast(num_blocks) * + sizeof( + tinyexr::tinyexr_int64); // sizeof(header) + sizeof(offsetTable) + + std::vector > data_list( + static_cast(num_blocks)); + std::vector channel_offset_list( + static_cast(exr_header->num_channels)); + + int pixel_data_size = 0; + size_t channel_offset = 0; + for (size_t c = 0; c < static_cast(exr_header->num_channels); c++) { + channel_offset_list[c] = channel_offset; + if (exr_header->requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + pixel_data_size += sizeof(unsigned short); + channel_offset += sizeof(unsigned short); + } else if (exr_header->requested_pixel_types[c] == + TINYEXR_PIXELTYPE_FLOAT) { + pixel_data_size += sizeof(float); + channel_offset += sizeof(float); + } else if (exr_header->requested_pixel_types[c] == TINYEXR_PIXELTYPE_UINT) { + pixel_data_size += sizeof(unsigned int); + channel_offset += sizeof(unsigned int); + } else { + TEXR_ASSERT(0); + } + } + +#if TINYEXR_USE_ZFP + tinyexr::ZFPCompressionParam zfp_compression_param; + + // Use ZFP compression parameter from custom attributes(if such a parameter + // exists) + { + bool ret = tinyexr::FindZFPCompressionParam( + &zfp_compression_param, exr_header->custom_attributes, + exr_header->num_custom_attributes); + + if (!ret) { + // Use predefined compression parameter. + zfp_compression_param.type = 0; + zfp_compression_param.rate = 2; + } + } +#endif + + // TOOD(LTE): C++11 thread + +// Use signed int since some OpenMP compiler doesn't allow unsigned type for +// `parallel for` +#if TINYEXR_USE_OPENMP +#pragma omp parallel for +#endif + for (int i = 0; i < num_blocks; i++) { + size_t ii = static_cast(i); + int start_y = num_scanlines * i; + int endY = (std::min)(num_scanlines * (i + 1), exr_image->height); + int h = endY - start_y; + + std::vector buf( + static_cast(exr_image->width * h * pixel_data_size)); + + for (size_t c = 0; c < static_cast(exr_header->num_channels); c++) { + if (exr_header->pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + if (exr_header->requested_pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT) { + for (int y = 0; y < h; y++) { + // Assume increasing Y + float *line_ptr = reinterpret_cast(&buf.at( + static_cast(pixel_data_size * y * exr_image->width) + + channel_offset_list[c] * + static_cast(exr_image->width))); + for (int x = 0; x < exr_image->width; x++) { + tinyexr::FP16 h16; + h16.u = reinterpret_cast( + exr_image->images)[c][(y + start_y) * exr_image->width + x]; + + tinyexr::FP32 f32 = half_to_float(h16); + + tinyexr::swap4(reinterpret_cast(&f32.f)); + + // line_ptr[x] = f32.f; + tinyexr::cpy4(line_ptr + x, &(f32.f)); + } + } + } else if (exr_header->requested_pixel_types[c] == + TINYEXR_PIXELTYPE_HALF) { + for (int y = 0; y < h; y++) { + // Assume increasing Y + unsigned short *line_ptr = reinterpret_cast( + &buf.at(static_cast(pixel_data_size * y * + exr_image->width) + + channel_offset_list[c] * + static_cast(exr_image->width))); + for (int x = 0; x < exr_image->width; x++) { + unsigned short val = reinterpret_cast( + exr_image->images)[c][(y + start_y) * exr_image->width + x]; + + tinyexr::swap2(&val); + + // line_ptr[x] = val; + tinyexr::cpy2(line_ptr + x, &val); + } + } + } else { + TEXR_ASSERT(0); + } + + } else if (exr_header->pixel_types[c] == TINYEXR_PIXELTYPE_FLOAT) { + if (exr_header->requested_pixel_types[c] == TINYEXR_PIXELTYPE_HALF) { + for (int y = 0; y < h; y++) { + // Assume increasing Y + unsigned short *line_ptr = reinterpret_cast( + &buf.at(static_cast(pixel_data_size * y * + exr_image->width) + + channel_offset_list[c] * + static_cast(exr_image->width))); + for (int x = 0; x < exr_image->width; x++) { + tinyexr::FP32 f32; + f32.f = reinterpret_cast( + exr_image->images)[c][(y + start_y) * exr_image->width + x]; + + tinyexr::FP16 h16; + h16 = float_to_half_full(f32); + + tinyexr::swap2(reinterpret_cast(&h16.u)); + + // line_ptr[x] = h16.u; + tinyexr::cpy2(line_ptr + x, &(h16.u)); + } + } + } else if (exr_header->requested_pixel_types[c] == + TINYEXR_PIXELTYPE_FLOAT) { + for (int y = 0; y < h; y++) { + // Assume increasing Y + float *line_ptr = reinterpret_cast(&buf.at( + static_cast(pixel_data_size * y * exr_image->width) + + channel_offset_list[c] * + static_cast(exr_image->width))); + for (int x = 0; x < exr_image->width; x++) { + float val = reinterpret_cast( + exr_image->images)[c][(y + start_y) * exr_image->width + x]; + + tinyexr::swap4(reinterpret_cast(&val)); + + // line_ptr[x] = val; + tinyexr::cpy4(line_ptr + x, &val); + } + } + } else { + TEXR_ASSERT(0); + } + } else if (exr_header->pixel_types[c] == TINYEXR_PIXELTYPE_UINT) { + for (int y = 0; y < h; y++) { + // Assume increasing Y + unsigned int *line_ptr = reinterpret_cast(&buf.at( + static_cast(pixel_data_size * y * exr_image->width) + + channel_offset_list[c] * static_cast(exr_image->width))); + for (int x = 0; x < exr_image->width; x++) { + unsigned int val = reinterpret_cast( + exr_image->images)[c][(y + start_y) * exr_image->width + x]; + + tinyexr::swap4(&val); + + // line_ptr[x] = val; + tinyexr::cpy4(line_ptr + x, &val); + } + } + } + } + + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_NONE) { + // 4 byte: scan line + // 4 byte: data size + // ~ : pixel data(uncompressed) + std::vector header(8); + unsigned int data_len = static_cast(buf.size()); + memcpy(&header.at(0), &start_y, sizeof(int)); + memcpy(&header.at(4), &data_len, sizeof(unsigned int)); + + tinyexr::swap4(reinterpret_cast(&header.at(0))); + tinyexr::swap4(reinterpret_cast(&header.at(4))); + + data_list[ii].insert(data_list[ii].end(), header.begin(), header.end()); + data_list[ii].insert(data_list[ii].end(), buf.begin(), + buf.begin() + data_len); + + } else if ((exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZIPS) || + (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZIP)) { +#if TINYEXR_USE_MINIZ + std::vector block(tinyexr::miniz::mz_compressBound( + static_cast(buf.size()))); +#else + std::vector block( + compressBound(static_cast(buf.size()))); +#endif + tinyexr::tinyexr_uint64 outSize = block.size(); + + tinyexr::CompressZip(&block.at(0), outSize, + reinterpret_cast(&buf.at(0)), + static_cast(buf.size())); + + // 4 byte: scan line + // 4 byte: data size + // ~ : pixel data(compressed) + std::vector header(8); + unsigned int data_len = static_cast(outSize); // truncate + memcpy(&header.at(0), &start_y, sizeof(int)); + memcpy(&header.at(4), &data_len, sizeof(unsigned int)); + + tinyexr::swap4(reinterpret_cast(&header.at(0))); + tinyexr::swap4(reinterpret_cast(&header.at(4))); + + data_list[ii].insert(data_list[ii].end(), header.begin(), header.end()); + data_list[ii].insert(data_list[ii].end(), block.begin(), + block.begin() + data_len); + + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_RLE) { + // (buf.size() * 3) / 2 would be enough. + std::vector block((buf.size() * 3) / 2); + + tinyexr::tinyexr_uint64 outSize = block.size(); + + tinyexr::CompressRle(&block.at(0), outSize, + reinterpret_cast(&buf.at(0)), + static_cast(buf.size())); + + // 4 byte: scan line + // 4 byte: data size + // ~ : pixel data(compressed) + std::vector header(8); + unsigned int data_len = static_cast(outSize); // truncate + memcpy(&header.at(0), &start_y, sizeof(int)); + memcpy(&header.at(4), &data_len, sizeof(unsigned int)); + + tinyexr::swap4(reinterpret_cast(&header.at(0))); + tinyexr::swap4(reinterpret_cast(&header.at(4))); + + data_list[ii].insert(data_list[ii].end(), header.begin(), header.end()); + data_list[ii].insert(data_list[ii].end(), block.begin(), + block.begin() + data_len); + + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { +#if TINYEXR_USE_PIZ + unsigned int bufLen = + 8192 + static_cast( + 2 * static_cast( + buf.size())); // @fixme { compute good bound. } + std::vector block(bufLen); + unsigned int outSize = static_cast(block.size()); + + CompressPiz(&block.at(0), &outSize, + reinterpret_cast(&buf.at(0)), + buf.size(), channels, exr_image->width, h); + + // 4 byte: scan line + // 4 byte: data size + // ~ : pixel data(compressed) + std::vector header(8); + unsigned int data_len = outSize; + memcpy(&header.at(0), &start_y, sizeof(int)); + memcpy(&header.at(4), &data_len, sizeof(unsigned int)); + + tinyexr::swap4(reinterpret_cast(&header.at(0))); + tinyexr::swap4(reinterpret_cast(&header.at(4))); + + data_list[ii].insert(data_list[ii].end(), header.begin(), header.end()); + data_list[ii].insert(data_list[ii].end(), block.begin(), + block.begin() + data_len); + +#else + TEXR_ASSERT(0); +#endif + } else if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { +#if TINYEXR_USE_ZFP + std::vector block; + unsigned int outSize; + + tinyexr::CompressZfp( + &block, &outSize, reinterpret_cast(&buf.at(0)), + exr_image->width, h, exr_header->num_channels, zfp_compression_param); + + // 4 byte: scan line + // 4 byte: data size + // ~ : pixel data(compressed) + std::vector header(8); + unsigned int data_len = outSize; + memcpy(&header.at(0), &start_y, sizeof(int)); + memcpy(&header.at(4), &data_len, sizeof(unsigned int)); + + tinyexr::swap4(reinterpret_cast(&header.at(0))); + tinyexr::swap4(reinterpret_cast(&header.at(4))); + + data_list[ii].insert(data_list[ii].end(), header.begin(), header.end()); + data_list[ii].insert(data_list[ii].end(), block.begin(), + block.begin() + data_len); + +#else + TEXR_ASSERT(0); +#endif + } else { + TEXR_ASSERT(0); + } + } // omp parallel + + for (size_t i = 0; i < static_cast(num_blocks); i++) { + offsets[i] = offset; + tinyexr::swap8(reinterpret_cast(&offsets[i])); + offset += data_list[i].size(); + } + + size_t totalSize = static_cast(offset); + { + memory.insert( + memory.end(), reinterpret_cast(&offsets.at(0)), + reinterpret_cast(&offsets.at(0)) + + sizeof(tinyexr::tinyexr_uint64) * static_cast(num_blocks)); + } + + if (memory.size() == 0) { + tinyexr::SetErrorMessage("Output memory size is zero", err); + return 0; + } + + (*memory_out) = static_cast(malloc(totalSize)); + memcpy((*memory_out), &memory.at(0), memory.size()); + unsigned char *memory_ptr = *memory_out + memory.size(); + + for (size_t i = 0; i < static_cast(num_blocks); i++) { + memcpy(memory_ptr, &data_list[i].at(0), data_list[i].size()); + memory_ptr += data_list[i].size(); + } + + return totalSize; // OK +} + +int SaveEXRImageToFile(const EXRImage *exr_image, const EXRHeader *exr_header, + const char *filename, const char **err) { + if (exr_image == NULL || filename == NULL || + exr_header->compression_type < 0) { + tinyexr::SetErrorMessage("Invalid argument for SaveEXRImageToFile", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#if !TINYEXR_USE_PIZ + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_PIZ) { + tinyexr::SetErrorMessage("PIZ compression is not supported in this build", + err); + return TINYEXR_ERROR_UNSUPPORTED_FEATURE; + } +#endif + +#if !TINYEXR_USE_ZFP + if (exr_header->compression_type == TINYEXR_COMPRESSIONTYPE_ZFP) { + tinyexr::SetErrorMessage("ZFP compression is not supported in this build", + err); + return TINYEXR_ERROR_UNSUPPORTED_FEATURE; + } +#endif + +#ifdef _WIN32 + FILE *fp = NULL; + fopen_s(&fp, filename, "wb"); +#else + FILE *fp = fopen(filename, "wb"); +#endif + if (!fp) { + tinyexr::SetErrorMessage("Cannot write a file", err); + return TINYEXR_ERROR_CANT_WRITE_FILE; + } + + unsigned char *mem = NULL; + size_t mem_size = SaveEXRImageToMemory(exr_image, exr_header, &mem, err); + if (mem_size == 0) { + fclose(fp); + return TINYEXR_ERROR_SERIALZATION_FAILED; + } + + size_t written_size = 0; + if ((mem_size > 0) && mem) { + written_size = fwrite(mem, 1, mem_size, fp); + } + free(mem); + + fclose(fp); + + if (written_size != mem_size) { + tinyexr::SetErrorMessage("Cannot write a file", err); + return TINYEXR_ERROR_CANT_WRITE_FILE; + } + + return TINYEXR_SUCCESS; +} + +int LoadDeepEXR(DeepImage *deep_image, const char *filename, const char **err) { + if (deep_image == NULL) { + tinyexr::SetErrorMessage("Invalid argument for LoadDeepEXR", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#ifdef _MSC_VER + FILE *fp = NULL; + errno_t errcode = fopen_s(&fp, filename, "rb"); + if ((0 != errcode) || (!fp)) { + tinyexr::SetErrorMessage("Cannot read a file " + std::string(filename), + err); + return TINYEXR_ERROR_CANT_OPEN_FILE; + } +#else + FILE *fp = fopen(filename, "rb"); + if (!fp) { + tinyexr::SetErrorMessage("Cannot read a file " + std::string(filename), + err); + return TINYEXR_ERROR_CANT_OPEN_FILE; + } +#endif + + size_t filesize; + // Compute size + fseek(fp, 0, SEEK_END); + filesize = static_cast(ftell(fp)); + fseek(fp, 0, SEEK_SET); + + if (filesize == 0) { + fclose(fp); + tinyexr::SetErrorMessage("File size is zero : " + std::string(filename), + err); + return TINYEXR_ERROR_INVALID_FILE; + } + + std::vector buf(filesize); // @todo { use mmap } + { + size_t ret; + ret = fread(&buf[0], 1, filesize, fp); + TEXR_ASSERT(ret == filesize); + (void)ret; + } + fclose(fp); + + const char *head = &buf[0]; + const char *marker = &buf[0]; + + // Header check. + { + const char header[] = {0x76, 0x2f, 0x31, 0x01}; + + if (memcmp(marker, header, 4) != 0) { + tinyexr::SetErrorMessage("Invalid magic number", err); + return TINYEXR_ERROR_INVALID_MAGIC_NUMBER; + } + marker += 4; + } + + // Version, scanline. + { + // ver 2.0, scanline, deep bit on(0x800) + // must be [2, 0, 0, 0] + if (marker[0] != 2 || marker[1] != 8 || marker[2] != 0 || marker[3] != 0) { + tinyexr::SetErrorMessage("Unsupported version or scanline", err); + return TINYEXR_ERROR_UNSUPPORTED_FORMAT; + } + + marker += 4; + } + + int dx = -1; + int dy = -1; + int dw = -1; + int dh = -1; + int num_scanline_blocks = 1; // 16 for ZIP compression. + int compression_type = -1; + int num_channels = -1; + std::vector channels; + + // Read attributes + size_t size = filesize - tinyexr::kEXRVersionSize; + for (;;) { + if (0 == size) { + return TINYEXR_ERROR_INVALID_DATA; + } else if (marker[0] == '\0') { + marker++; + size--; + break; + } + + std::string attr_name; + std::string attr_type; + std::vector data; + size_t marker_size; + if (!tinyexr::ReadAttribute(&attr_name, &attr_type, &data, &marker_size, + marker, size)) { + std::stringstream ss; + ss << "Failed to parse attribute\n"; + tinyexr::SetErrorMessage(ss.str(), err); + return TINYEXR_ERROR_INVALID_DATA; + } + marker += marker_size; + size -= marker_size; + + if (attr_name.compare("compression") == 0) { + compression_type = data[0]; + if (compression_type > TINYEXR_COMPRESSIONTYPE_PIZ) { + std::stringstream ss; + ss << "Unsupported compression type : " << compression_type; + tinyexr::SetErrorMessage(ss.str(), err); + return TINYEXR_ERROR_UNSUPPORTED_FORMAT; + } + + if (compression_type == TINYEXR_COMPRESSIONTYPE_ZIP) { + num_scanline_blocks = 16; + } + + } else if (attr_name.compare("channels") == 0) { + // name: zero-terminated string, from 1 to 255 bytes long + // pixel type: int, possible values are: UINT = 0 HALF = 1 FLOAT = 2 + // pLinear: unsigned char, possible values are 0 and 1 + // reserved: three chars, should be zero + // xSampling: int + // ySampling: int + + if (!tinyexr::ReadChannelInfo(channels, data)) { + tinyexr::SetErrorMessage("Failed to parse channel info", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + num_channels = static_cast(channels.size()); + + if (num_channels < 1) { + tinyexr::SetErrorMessage("Invalid channels format", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + } else if (attr_name.compare("dataWindow") == 0) { + memcpy(&dx, &data.at(0), sizeof(int)); + memcpy(&dy, &data.at(4), sizeof(int)); + memcpy(&dw, &data.at(8), sizeof(int)); + memcpy(&dh, &data.at(12), sizeof(int)); + tinyexr::swap4(reinterpret_cast(&dx)); + tinyexr::swap4(reinterpret_cast(&dy)); + tinyexr::swap4(reinterpret_cast(&dw)); + tinyexr::swap4(reinterpret_cast(&dh)); + + } else if (attr_name.compare("displayWindow") == 0) { + int x; + int y; + int w; + int h; + memcpy(&x, &data.at(0), sizeof(int)); + memcpy(&y, &data.at(4), sizeof(int)); + memcpy(&w, &data.at(8), sizeof(int)); + memcpy(&h, &data.at(12), sizeof(int)); + tinyexr::swap4(reinterpret_cast(&x)); + tinyexr::swap4(reinterpret_cast(&y)); + tinyexr::swap4(reinterpret_cast(&w)); + tinyexr::swap4(reinterpret_cast(&h)); + } + } + + TEXR_ASSERT(dx >= 0); + TEXR_ASSERT(dy >= 0); + TEXR_ASSERT(dw >= 0); + TEXR_ASSERT(dh >= 0); + TEXR_ASSERT(num_channels >= 1); + + int data_width = dw - dx + 1; + int data_height = dh - dy + 1; + + std::vector image( + static_cast(data_width * data_height * 4)); // 4 = RGBA + + // Read offset tables. + int num_blocks = data_height / num_scanline_blocks; + if (num_blocks * num_scanline_blocks < data_height) { + num_blocks++; + } + + std::vector offsets(static_cast(num_blocks)); + + for (size_t y = 0; y < static_cast(num_blocks); y++) { + tinyexr::tinyexr_int64 offset; + memcpy(&offset, marker, sizeof(tinyexr::tinyexr_int64)); + tinyexr::swap8(reinterpret_cast(&offset)); + marker += sizeof(tinyexr::tinyexr_int64); // = 8 + offsets[y] = offset; + } + +#if TINYEXR_USE_PIZ + if ((compression_type == TINYEXR_COMPRESSIONTYPE_NONE) || + (compression_type == TINYEXR_COMPRESSIONTYPE_RLE) || + (compression_type == TINYEXR_COMPRESSIONTYPE_ZIPS) || + (compression_type == TINYEXR_COMPRESSIONTYPE_ZIP) || + (compression_type == TINYEXR_COMPRESSIONTYPE_PIZ)) { +#else + if ((compression_type == TINYEXR_COMPRESSIONTYPE_NONE) || + (compression_type == TINYEXR_COMPRESSIONTYPE_RLE) || + (compression_type == TINYEXR_COMPRESSIONTYPE_ZIPS) || + (compression_type == TINYEXR_COMPRESSIONTYPE_ZIP)) { +#endif + // OK + } else { + tinyexr::SetErrorMessage("Unsupported compression format", err); + return TINYEXR_ERROR_UNSUPPORTED_FORMAT; + } + + deep_image->image = static_cast( + malloc(sizeof(float **) * static_cast(num_channels))); + for (int c = 0; c < num_channels; c++) { + deep_image->image[c] = static_cast( + malloc(sizeof(float *) * static_cast(data_height))); + for (int y = 0; y < data_height; y++) { + } + } + + deep_image->offset_table = static_cast( + malloc(sizeof(int *) * static_cast(data_height))); + for (int y = 0; y < data_height; y++) { + deep_image->offset_table[y] = static_cast( + malloc(sizeof(int) * static_cast(data_width))); + } + + for (size_t y = 0; y < static_cast(num_blocks); y++) { + const unsigned char *data_ptr = + reinterpret_cast(head + offsets[y]); + + // int: y coordinate + // int64: packed size of pixel offset table + // int64: packed size of sample data + // int64: unpacked size of sample data + // compressed pixel offset table + // compressed sample data + int line_no; + tinyexr::tinyexr_int64 packedOffsetTableSize; + tinyexr::tinyexr_int64 packedSampleDataSize; + tinyexr::tinyexr_int64 unpackedSampleDataSize; + memcpy(&line_no, data_ptr, sizeof(int)); + memcpy(&packedOffsetTableSize, data_ptr + 4, + sizeof(tinyexr::tinyexr_int64)); + memcpy(&packedSampleDataSize, data_ptr + 12, + sizeof(tinyexr::tinyexr_int64)); + memcpy(&unpackedSampleDataSize, data_ptr + 20, + sizeof(tinyexr::tinyexr_int64)); + + tinyexr::swap4(reinterpret_cast(&line_no)); + tinyexr::swap8( + reinterpret_cast(&packedOffsetTableSize)); + tinyexr::swap8( + reinterpret_cast(&packedSampleDataSize)); + tinyexr::swap8( + reinterpret_cast(&unpackedSampleDataSize)); + + std::vector pixelOffsetTable(static_cast(data_width)); + + // decode pixel offset table. + { + unsigned long dstLen = + static_cast(pixelOffsetTable.size() * sizeof(int)); + if (!tinyexr::DecompressZip( + reinterpret_cast(&pixelOffsetTable.at(0)), + &dstLen, data_ptr + 28, + static_cast(packedOffsetTableSize))) { + return false; + } + + TEXR_ASSERT(dstLen == pixelOffsetTable.size() * sizeof(int)); + for (size_t i = 0; i < static_cast(data_width); i++) { + deep_image->offset_table[y][i] = pixelOffsetTable[i]; + } + } + + std::vector sample_data( + static_cast(unpackedSampleDataSize)); + + // decode sample data. + { + unsigned long dstLen = static_cast(unpackedSampleDataSize); + if (dstLen) { + if (!tinyexr::DecompressZip( + reinterpret_cast(&sample_data.at(0)), &dstLen, + data_ptr + 28 + packedOffsetTableSize, + static_cast(packedSampleDataSize))) { + return false; + } + TEXR_ASSERT(dstLen == static_cast(unpackedSampleDataSize)); + } + } + + // decode sample + int sampleSize = -1; + std::vector channel_offset_list(static_cast(num_channels)); + { + int channel_offset = 0; + for (size_t i = 0; i < static_cast(num_channels); i++) { + channel_offset_list[i] = channel_offset; + if (channels[i].pixel_type == TINYEXR_PIXELTYPE_UINT) { // UINT + channel_offset += 4; + } else if (channels[i].pixel_type == TINYEXR_PIXELTYPE_HALF) { // half + channel_offset += 2; + } else if (channels[i].pixel_type == + TINYEXR_PIXELTYPE_FLOAT) { // float + channel_offset += 4; + } else { + TEXR_ASSERT(0); + } + } + sampleSize = channel_offset; + } + TEXR_ASSERT(sampleSize >= 2); + + TEXR_ASSERT(static_cast( + pixelOffsetTable[static_cast(data_width - 1)] * + sampleSize) == sample_data.size()); + int samples_per_line = static_cast(sample_data.size()) / sampleSize; + + // + // Alloc memory + // + + // + // pixel data is stored as image[channels][pixel_samples] + // + { + tinyexr::tinyexr_uint64 data_offset = 0; + for (size_t c = 0; c < static_cast(num_channels); c++) { + deep_image->image[c][y] = static_cast( + malloc(sizeof(float) * static_cast(samples_per_line))); + + if (channels[c].pixel_type == 0) { // UINT + for (size_t x = 0; x < static_cast(samples_per_line); x++) { + unsigned int ui; + unsigned int *src_ptr = reinterpret_cast( + &sample_data.at(size_t(data_offset) + x * sizeof(int))); + tinyexr::cpy4(&ui, src_ptr); + deep_image->image[c][y][x] = static_cast(ui); // @fixme + } + data_offset += + sizeof(unsigned int) * static_cast(samples_per_line); + } else if (channels[c].pixel_type == 1) { // half + for (size_t x = 0; x < static_cast(samples_per_line); x++) { + tinyexr::FP16 f16; + const unsigned short *src_ptr = reinterpret_cast( + &sample_data.at(size_t(data_offset) + x * sizeof(short))); + tinyexr::cpy2(&(f16.u), src_ptr); + tinyexr::FP32 f32 = half_to_float(f16); + deep_image->image[c][y][x] = f32.f; + } + data_offset += sizeof(short) * static_cast(samples_per_line); + } else { // float + for (size_t x = 0; x < static_cast(samples_per_line); x++) { + float f; + const float *src_ptr = reinterpret_cast( + &sample_data.at(size_t(data_offset) + x * sizeof(float))); + tinyexr::cpy4(&f, src_ptr); + deep_image->image[c][y][x] = f; + } + data_offset += sizeof(float) * static_cast(samples_per_line); + } + } + } + } // y + + deep_image->width = data_width; + deep_image->height = data_height; + + deep_image->channel_names = static_cast( + malloc(sizeof(const char *) * static_cast(num_channels))); + for (size_t c = 0; c < static_cast(num_channels); c++) { +#ifdef _WIN32 + deep_image->channel_names[c] = _strdup(channels[c].name.c_str()); +#else + deep_image->channel_names[c] = strdup(channels[c].name.c_str()); +#endif + } + deep_image->num_channels = num_channels; + + return TINYEXR_SUCCESS; +} + +void InitEXRImage(EXRImage *exr_image) { + if (exr_image == NULL) { + return; + } + + exr_image->width = 0; + exr_image->height = 0; + exr_image->num_channels = 0; + + exr_image->images = NULL; + exr_image->tiles = NULL; + + exr_image->num_tiles = 0; +} + +void FreeEXRErrorMessage(const char *msg) { + if (msg) { + free(reinterpret_cast(const_cast(msg))); + } + return; +} + +void InitEXRHeader(EXRHeader *exr_header) { + if (exr_header == NULL) { + return; + } + + memset(exr_header, 0, sizeof(EXRHeader)); +} + +int FreeEXRHeader(EXRHeader *exr_header) { + if (exr_header == NULL) { + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + if (exr_header->channels) { + free(exr_header->channels); + } + + if (exr_header->pixel_types) { + free(exr_header->pixel_types); + } + + if (exr_header->requested_pixel_types) { + free(exr_header->requested_pixel_types); + } + + for (int i = 0; i < exr_header->num_custom_attributes; i++) { + if (exr_header->custom_attributes[i].value) { + free(exr_header->custom_attributes[i].value); + } + } + + if (exr_header->custom_attributes) { + free(exr_header->custom_attributes); + } + + return TINYEXR_SUCCESS; +} + +int FreeEXRImage(EXRImage *exr_image) { + if (exr_image == NULL) { + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + for (int i = 0; i < exr_image->num_channels; i++) { + if (exr_image->images && exr_image->images[i]) { + free(exr_image->images[i]); + } + } + + if (exr_image->images) { + free(exr_image->images); + } + + if (exr_image->tiles) { + for (int tid = 0; tid < exr_image->num_tiles; tid++) { + for (int i = 0; i < exr_image->num_channels; i++) { + if (exr_image->tiles[tid].images && exr_image->tiles[tid].images[i]) { + free(exr_image->tiles[tid].images[i]); + } + } + if (exr_image->tiles[tid].images) { + free(exr_image->tiles[tid].images); + } + } + free(exr_image->tiles); + } + + return TINYEXR_SUCCESS; +} + +int ParseEXRHeaderFromFile(EXRHeader *exr_header, const EXRVersion *exr_version, + const char *filename, const char **err) { + if (exr_header == NULL || exr_version == NULL || filename == NULL) { + tinyexr::SetErrorMessage("Invalid argument for ParseEXRHeaderFromFile", + err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#ifdef _WIN32 + FILE *fp = NULL; + fopen_s(&fp, filename, "rb"); +#else + FILE *fp = fopen(filename, "rb"); +#endif + if (!fp) { + tinyexr::SetErrorMessage("Cannot read file " + std::string(filename), err); + return TINYEXR_ERROR_CANT_OPEN_FILE; + } + + size_t filesize; + // Compute size + fseek(fp, 0, SEEK_END); + filesize = static_cast(ftell(fp)); + fseek(fp, 0, SEEK_SET); + + std::vector buf(filesize); // @todo { use mmap } + { + size_t ret; + ret = fread(&buf[0], 1, filesize, fp); + TEXR_ASSERT(ret == filesize); + fclose(fp); + + if (ret != filesize) { + tinyexr::SetErrorMessage("fread() error on " + std::string(filename), + err); + return TINYEXR_ERROR_INVALID_FILE; + } + } + + return ParseEXRHeaderFromMemory(exr_header, exr_version, &buf.at(0), filesize, + err); +} + +int ParseEXRMultipartHeaderFromMemory(EXRHeader ***exr_headers, + int *num_headers, + const EXRVersion *exr_version, + const unsigned char *memory, size_t size, + const char **err) { + if (memory == NULL || exr_headers == NULL || num_headers == NULL || + exr_version == NULL) { + // Invalid argument + tinyexr::SetErrorMessage( + "Invalid argument for ParseEXRMultipartHeaderFromMemory", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + if (size < tinyexr::kEXRVersionSize) { + tinyexr::SetErrorMessage("Data size too short", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + const unsigned char *marker = memory + tinyexr::kEXRVersionSize; + size_t marker_size = size - tinyexr::kEXRVersionSize; + + std::vector infos; + + for (;;) { + tinyexr::HeaderInfo info; + info.clear(); + + std::string err_str; + bool empty_header = false; + int ret = ParseEXRHeader(&info, &empty_header, exr_version, &err_str, + marker, marker_size); + + if (ret != TINYEXR_SUCCESS) { + tinyexr::SetErrorMessage(err_str, err); + return ret; + } + + if (empty_header) { + marker += 1; // skip '\0' + break; + } + + // `chunkCount` must exist in the header. + if (info.chunk_count == 0) { + tinyexr::SetErrorMessage( + "`chunkCount' attribute is not found in the header.", err); + return TINYEXR_ERROR_INVALID_DATA; + } + + infos.push_back(info); + + // move to next header. + marker += info.header_len; + size -= info.header_len; + } + + // allocate memory for EXRHeader and create array of EXRHeader pointers. + (*exr_headers) = + static_cast(malloc(sizeof(EXRHeader *) * infos.size())); + for (size_t i = 0; i < infos.size(); i++) { + EXRHeader *exr_header = static_cast(malloc(sizeof(EXRHeader))); + + ConvertHeader(exr_header, infos[i]); + + // transfoer `tiled` from version. + exr_header->tiled = exr_version->tiled; + + (*exr_headers)[i] = exr_header; + } + + (*num_headers) = static_cast(infos.size()); + + return TINYEXR_SUCCESS; +} + +int ParseEXRMultipartHeaderFromFile(EXRHeader ***exr_headers, int *num_headers, + const EXRVersion *exr_version, + const char *filename, const char **err) { + if (exr_headers == NULL || num_headers == NULL || exr_version == NULL || + filename == NULL) { + tinyexr::SetErrorMessage( + "Invalid argument for ParseEXRMultipartHeaderFromFile()", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#ifdef _WIN32 + FILE *fp = NULL; + fopen_s(&fp, filename, "rb"); +#else + FILE *fp = fopen(filename, "rb"); +#endif + if (!fp) { + tinyexr::SetErrorMessage("Cannot read file " + std::string(filename), err); + return TINYEXR_ERROR_CANT_OPEN_FILE; + } + + size_t filesize; + // Compute size + fseek(fp, 0, SEEK_END); + filesize = static_cast(ftell(fp)); + fseek(fp, 0, SEEK_SET); + + std::vector buf(filesize); // @todo { use mmap } + { + size_t ret; + ret = fread(&buf[0], 1, filesize, fp); + TEXR_ASSERT(ret == filesize); + fclose(fp); + + if (ret != filesize) { + tinyexr::SetErrorMessage("`fread' error. file may be corrupted.", err); + return TINYEXR_ERROR_INVALID_FILE; + } + } + + return ParseEXRMultipartHeaderFromMemory( + exr_headers, num_headers, exr_version, &buf.at(0), filesize, err); +} + +int ParseEXRVersionFromMemory(EXRVersion *version, const unsigned char *memory, + size_t size) { + if (version == NULL || memory == NULL) { + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + if (size < tinyexr::kEXRVersionSize) { + return TINYEXR_ERROR_INVALID_DATA; + } + + const unsigned char *marker = memory; + + // Header check. + { + const char header[] = {0x76, 0x2f, 0x31, 0x01}; + + if (memcmp(marker, header, 4) != 0) { + return TINYEXR_ERROR_INVALID_MAGIC_NUMBER; + } + marker += 4; + } + + version->tiled = false; + version->long_name = false; + version->non_image = false; + version->multipart = false; + + // Parse version header. + { + // must be 2 + if (marker[0] != 2) { + return TINYEXR_ERROR_INVALID_EXR_VERSION; + } + + version->version = 2; + + if (marker[1] & 0x2) { // 9th bit + version->tiled = true; + } + if (marker[1] & 0x4) { // 10th bit + version->long_name = true; + } + if (marker[1] & 0x8) { // 11th bit + version->non_image = true; // (deep image) + } + if (marker[1] & 0x10) { // 12th bit + version->multipart = true; + } + } + + return TINYEXR_SUCCESS; +} + +int ParseEXRVersionFromFile(EXRVersion *version, const char *filename) { + if (filename == NULL) { + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#ifdef _WIN32 + FILE *fp = NULL; + fopen_s(&fp, filename, "rb"); +#else + FILE *fp = fopen(filename, "rb"); +#endif + if (!fp) { + return TINYEXR_ERROR_CANT_OPEN_FILE; + } + + size_t file_size; + // Compute size + fseek(fp, 0, SEEK_END); + file_size = static_cast(ftell(fp)); + fseek(fp, 0, SEEK_SET); + + if (file_size < tinyexr::kEXRVersionSize) { + fclose(fp); + return TINYEXR_ERROR_INVALID_FILE; + } + + unsigned char buf[tinyexr::kEXRVersionSize]; + size_t ret = fread(&buf[0], 1, tinyexr::kEXRVersionSize, fp); + fclose(fp); + + if (ret != tinyexr::kEXRVersionSize) { + return TINYEXR_ERROR_INVALID_FILE; + } + + return ParseEXRVersionFromMemory(version, buf, tinyexr::kEXRVersionSize); +} + +int LoadEXRMultipartImageFromMemory(EXRImage *exr_images, + const EXRHeader **exr_headers, + unsigned int num_parts, + const unsigned char *memory, + const size_t size, const char **err) { + if (exr_images == NULL || exr_headers == NULL || num_parts == 0 || + memory == NULL || (size <= tinyexr::kEXRVersionSize)) { + tinyexr::SetErrorMessage( + "Invalid argument for LoadEXRMultipartImageFromMemory()", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + // compute total header size. + size_t total_header_size = 0; + for (unsigned int i = 0; i < num_parts; i++) { + if (exr_headers[i]->header_len == 0) { + tinyexr::SetErrorMessage("EXRHeader variable is not initialized.", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + total_header_size += exr_headers[i]->header_len; + } + + const char *marker = reinterpret_cast( + memory + total_header_size + 4 + + 4); // +8 for magic number and version header. + + marker += 1; // Skip empty header. + + // NOTE 1: + // In multipart image, There is 'part number' before chunk data. + // 4 byte : part number + // 4+ : chunk + // + // NOTE 2: + // EXR spec says 'part number' is 'unsigned long' but actually this is + // 'unsigned int(4 bytes)' in OpenEXR implementation... + // http://www.openexr.com/openexrfilelayout.pdf + + // Load chunk offset table. + std::vector > chunk_offset_table_list; + for (size_t i = 0; i < static_cast(num_parts); i++) { + std::vector offset_table( + static_cast(exr_headers[i]->chunk_count)); + + for (size_t c = 0; c < offset_table.size(); c++) { + tinyexr::tinyexr_uint64 offset; + memcpy(&offset, marker, 8); + tinyexr::swap8(&offset); + + if (offset >= size) { + tinyexr::SetErrorMessage("Invalid offset size in EXR header chunks.", + err); + return TINYEXR_ERROR_INVALID_DATA; + } + + offset_table[c] = offset + 4; // +4 to skip 'part number' + marker += 8; + } + + chunk_offset_table_list.push_back(offset_table); + } + + // Decode image. + for (size_t i = 0; i < static_cast(num_parts); i++) { + std::vector &offset_table = + chunk_offset_table_list[i]; + + // First check 'part number' is identitical to 'i' + for (size_t c = 0; c < offset_table.size(); c++) { + const unsigned char *part_number_addr = + memory + offset_table[c] - 4; // -4 to move to 'part number' field. + unsigned int part_no; + memcpy(&part_no, part_number_addr, sizeof(unsigned int)); // 4 + tinyexr::swap4(&part_no); + + if (part_no != i) { + tinyexr::SetErrorMessage("Invalid `part number' in EXR header chunks.", + err); + return TINYEXR_ERROR_INVALID_DATA; + } + } + + std::string e; + int ret = tinyexr::DecodeChunk(&exr_images[i], exr_headers[i], offset_table, + memory, size, &e); + if (ret != TINYEXR_SUCCESS) { + if (!e.empty()) { + tinyexr::SetErrorMessage(e, err); + } + return ret; + } + } + + return TINYEXR_SUCCESS; +} + +int LoadEXRMultipartImageFromFile(EXRImage *exr_images, + const EXRHeader **exr_headers, + unsigned int num_parts, const char *filename, + const char **err) { + if (exr_images == NULL || exr_headers == NULL || num_parts == 0) { + tinyexr::SetErrorMessage( + "Invalid argument for LoadEXRMultipartImageFromFile", err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + +#ifdef _WIN32 + FILE *fp = NULL; + fopen_s(&fp, filename, "rb"); +#else + FILE *fp = fopen(filename, "rb"); +#endif + if (!fp) { + tinyexr::SetErrorMessage("Cannot read file " + std::string(filename), err); + return TINYEXR_ERROR_CANT_OPEN_FILE; + } + + size_t filesize; + // Compute size + fseek(fp, 0, SEEK_END); + filesize = static_cast(ftell(fp)); + fseek(fp, 0, SEEK_SET); + + std::vector buf(filesize); // @todo { use mmap } + { + size_t ret; + ret = fread(&buf[0], 1, filesize, fp); + TEXR_ASSERT(ret == filesize); + fclose(fp); + (void)ret; + } + + return LoadEXRMultipartImageFromMemory(exr_images, exr_headers, num_parts, + &buf.at(0), filesize, err); +} + +int SaveEXR(const float *data, int width, int height, int components, + const int save_as_fp16, const char *outfilename, const char **err) { + if ((components == 1) || components == 3 || components == 4) { + // OK + } else { + std::stringstream ss; + ss << "Unsupported component value : " << components << std::endl; + + tinyexr::SetErrorMessage(ss.str(), err); + return TINYEXR_ERROR_INVALID_ARGUMENT; + } + + EXRHeader header; + InitEXRHeader(&header); + + if ((width < 16) && (height < 16)) { + // No compression for small image. + header.compression_type = TINYEXR_COMPRESSIONTYPE_NONE; + } else { + header.compression_type = TINYEXR_COMPRESSIONTYPE_ZIP; + } + + EXRImage image; + InitEXRImage(&image); + + image.num_channels = components; + + std::vector images[4]; + + if (components == 1) { + images[0].resize(static_cast(width * height)); + memcpy(images[0].data(), data, sizeof(float) * size_t(width * height)); + } else { + images[0].resize(static_cast(width * height)); + images[1].resize(static_cast(width * height)); + images[2].resize(static_cast(width * height)); + images[3].resize(static_cast(width * height)); + + // Split RGB(A)RGB(A)RGB(A)... into R, G and B(and A) layers + for (size_t i = 0; i < static_cast(width * height); i++) { + images[0][i] = data[static_cast(components) * i + 0]; + images[1][i] = data[static_cast(components) * i + 1]; + images[2][i] = data[static_cast(components) * i + 2]; + if (components == 4) { + images[3][i] = data[static_cast(components) * i + 3]; + } + } + } + + float *image_ptr[4] = {0, 0, 0, 0}; + if (components == 4) { + image_ptr[0] = &(images[3].at(0)); // A + image_ptr[1] = &(images[2].at(0)); // B + image_ptr[2] = &(images[1].at(0)); // G + image_ptr[3] = &(images[0].at(0)); // R + } else if (components == 3) { + image_ptr[0] = &(images[2].at(0)); // B + image_ptr[1] = &(images[1].at(0)); // G + image_ptr[2] = &(images[0].at(0)); // R + } else if (components == 1) { + image_ptr[0] = &(images[0].at(0)); // A + } + + image.images = reinterpret_cast(image_ptr); + image.width = width; + image.height = height; + + header.num_channels = components; + header.channels = static_cast(malloc( + sizeof(EXRChannelInfo) * static_cast(header.num_channels))); + // Must be (A)BGR order, since most of EXR viewers expect this channel order. + if (components == 4) { +#ifdef _MSC_VER + strncpy_s(header.channels[0].name, "A", 255); + strncpy_s(header.channels[1].name, "B", 255); + strncpy_s(header.channels[2].name, "G", 255); + strncpy_s(header.channels[3].name, "R", 255); +#else + strncpy(header.channels[0].name, "A", 255); + strncpy(header.channels[1].name, "B", 255); + strncpy(header.channels[2].name, "G", 255); + strncpy(header.channels[3].name, "R", 255); +#endif + header.channels[0].name[strlen("A")] = '\0'; + header.channels[1].name[strlen("B")] = '\0'; + header.channels[2].name[strlen("G")] = '\0'; + header.channels[3].name[strlen("R")] = '\0'; + } else if (components == 3) { +#ifdef _MSC_VER + strncpy_s(header.channels[0].name, "B", 255); + strncpy_s(header.channels[1].name, "G", 255); + strncpy_s(header.channels[2].name, "R", 255); +#else + strncpy(header.channels[0].name, "B", 255); + strncpy(header.channels[1].name, "G", 255); + strncpy(header.channels[2].name, "R", 255); +#endif + header.channels[0].name[strlen("B")] = '\0'; + header.channels[1].name[strlen("G")] = '\0'; + header.channels[2].name[strlen("R")] = '\0'; + } else { +#ifdef _MSC_VER + strncpy_s(header.channels[0].name, "A", 255); +#else + strncpy(header.channels[0].name, "A", 255); +#endif + header.channels[0].name[strlen("A")] = '\0'; + } + + header.pixel_types = static_cast( + malloc(sizeof(int) * static_cast(header.num_channels))); + header.requested_pixel_types = static_cast( + malloc(sizeof(int) * static_cast(header.num_channels))); + for (int i = 0; i < header.num_channels; i++) { + header.pixel_types[i] = + TINYEXR_PIXELTYPE_FLOAT; // pixel type of input image + + if (save_as_fp16 > 0) { + header.requested_pixel_types[i] = + TINYEXR_PIXELTYPE_HALF; // save with half(fp16) pixel format + } else { + header.requested_pixel_types[i] = + TINYEXR_PIXELTYPE_FLOAT; // save with float(fp32) pixel format(i.e. + // no precision reduction) + } + } + + int ret = SaveEXRImageToFile(&image, &header, outfilename, err); + if (ret != TINYEXR_SUCCESS) { + return ret; + } + + free(header.channels); + free(header.pixel_types); + free(header.requested_pixel_types); + + return ret; +} + +#ifdef __clang__ +// zero-as-null-ppinter-constant +#pragma clang diagnostic pop +#endif + +#endif // TINYEXR_IMPLEMENTATION_DEIFNED +#endif // TINYEXR_IMPLEMENTATION diff --git a/tools/imageio/image.hpp b/tools/imageio/image.hpp index 4f4026d02e..6d4dac3801 100644 --- a/tools/imageio/image.hpp +++ b/tools/imageio/image.hpp @@ -37,8 +37,8 @@ #if defined(__GNUC__) && !defined(__clang__) #pragma GCC diagnostic ignored "-Wunused-value" #endif -#include "encoder/basisu_resampler.h" -#include "encoder/basisu_resampler_filters.h" +#include "basisu/encoder/basisu_resampler.h" +#include "basisu/encoder/basisu_resampler_filters.h" #if defined(__GCC__) && !defined(__clang__) #pragma GCC diagnostic pop #endif diff --git a/tools/imageio/jpg.imageio/jpginput.cc b/tools/imageio/jpg.imageio/jpginput.cc index 4bd4915405..b8f50cb246 100644 --- a/tools/imageio/jpg.imageio/jpginput.cc +++ b/tools/imageio/jpg.imageio/jpginput.cc @@ -26,7 +26,7 @@ #include #include -#include "encoder/jpgd.h" +#include "jpgd.h" using namespace jpgd; diff --git a/tools/ktx/CMakeLists.txt b/tools/ktx/CMakeLists.txt index 77bd31102f..56d4ba7757 100644 --- a/tools/ktx/CMakeLists.txt +++ b/tools/ktx/CMakeLists.txt @@ -63,8 +63,8 @@ PRIVATE imageio objUtil ktx - basisu_encoder - ${ASTCENC_LIB_TARGET} + basisu::basisu_encoder + astc::astcenc $,Pathcch,> # For PathCchRemoveFileSpec on Windows fmt::fmt cxxopts::cxxopts diff --git a/tools/ktx/command_extract.cpp b/tools/ktx/command_extract.cpp index 5d11bd70ec..47a2a98652 100644 --- a/tools/ktx/command_extract.cpp +++ b/tools/ktx/command_extract.cpp @@ -23,9 +23,9 @@ #include #include -#include "lodepng/lodepng.h" -#include "astc-encoder/Source/ThirdParty/tinyexr.h" -#include "astc-encoder/Source/astcenc.h" +#include "lodepng.h" +#include "tinyexr.h" +#include // ------------------------------------------------------------------------------------------------- diff --git a/tools/ktx/metrics_utils.h b/tools/ktx/metrics_utils.h index 0d4eded9de..b3aace4877 100644 --- a/tools/ktx/metrics_utils.h +++ b/tools/ktx/metrics_utils.h @@ -13,8 +13,8 @@ #include "ktxint.h" #include "texture2.h" -#include -#include +#include "basisu/encoder/basisu_enc.h" +#include "basisu/encoder/basisu_ssim.h" // ------------------------------------------------------------------------------------------------- From 4e6284f62f31a5edce883c7ce12ca5e8375337b6 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Sun, 14 Dec 2025 20:35:34 +0200 Subject: [PATCH 02/20] fixed dependency linking for installed target --- external/basisu.cmake | 11 ++++++++--- lib/CMakeLists.txt | 12 +++++------- 2 files changed, 13 insertions(+), 10 deletions(-) diff --git a/external/basisu.cmake b/external/basisu.cmake index 682357a836..6ec6aa7a79 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -54,7 +54,12 @@ if (NOT TARGET basisu::basisu_encoder) # Iterating over collected files foreach(filename ${headers}) get_filename_component(directory ${filename} DIRECTORY) - file(${BASISU_INCLUDE_DIR} ${BASISU_INCLUDE_DIR}/${subdir}/${directory}) + + # Ensure destination directory exists + file( + MAKE_DIRECTORY + ${BASISU_INCLUDE_DIR} ${BASISU_INCLUDE_DIR}/${subdir}/${directory} + ) # And finally copy to temporary folder file(COPY ${basisu_SOURCE_DIR}/${subdir}/${filename} @@ -63,7 +68,7 @@ if (NOT TARGET basisu::basisu_encoder) endforeach() # Add temporary include path to target - target_include_directories(basisu_encoder - PUBLIC ${CMAKE_CURRENT_BINARY_DIR}/include + target_include_directories(basisu_encoder PUBLIC + $ ) endif() diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index bb48786135..48d8f2cd2d 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -475,12 +475,7 @@ macro(common_libktx_settings target enable_write library_type) target_link_libraries( ${target} PRIVATE - zstd::libzstd - ) - - target_link_libraries( - ${target} PRIVATE - astc::astcenc + $ ) endmacro(common_libktx_settings) @@ -559,7 +554,10 @@ if(LIBKTX_VERSION_READ_ONLY AND NOT (BUILD_SHARED_LIBRARIES OR LIBKTX_VERSION_FU list(APPEND LIBKTX_INSTALL_TARGETS ktx_read) endif() if(NOT BUILD_SHARED_LIBS AND NOT APPLE) - list(APPEND LIBKTX_INSTALL_TARGETS astc::astcenc basisu::basisu_encoder) + get_target_property(ASTC_LIB_NAME astc::astcenc ALIASED_TARGET) + get_target_property(BASISU_LIB_NAME basisu::basisu_encoder ALIASED_TARGET) + + list(APPEND LIBKTX_INSTALL_TARGETS ${ASTC_LIB_NAME} ${BASISU_LIB_NAME}) endif() if(APPLE OR LINUX) From 3751c9e30f3ff11e811c21ea2aa8b32ad18875b9 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Sun, 14 Dec 2025 22:46:13 +0200 Subject: [PATCH 03/20] removed linker optimization macro --- lib/CMakeLists.txt | 17 ++++++----------- 1 file changed, 6 insertions(+), 11 deletions(-) diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index 48d8f2cd2d..f0acadce75 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -479,15 +479,6 @@ macro(common_libktx_settings target enable_write library_type) ) endmacro(common_libktx_settings) -macro(add_lib_dependencies target lib) - if(NOT BUILD_SHARED_LIBS) - # Make a single static library to simplify linking. - target_sources(${target} PRIVATE $) - target_include_directories(${target} PRIVATE $) - else() - target_link_libraries(${target} PRIVATE ${lib}) - endif() -endmacro(add_lib_dependencies) ################################################### # ktx libraries @@ -501,7 +492,9 @@ if(LIBKTX_VERSION_FULL) src/writer2.c ) common_libktx_settings(ktx 1 ${LIB_TYPE}) - add_lib_dependencies(ktx basisu::basisu_encoder) + target_link_libraries( + ktx PRIVATE basisu::basisu_encoder + ) endif() # Read-only library @@ -510,7 +503,9 @@ if(LIBKTX_VERSION_READ_ONLY) ${LIBKTX_MAIN_SRC} ) common_libktx_settings(ktx_read 0 ${LIB_TYPE}) - add_lib_dependencies(ktx_read basisu::basisu_encoder) + target_link_libraries( + ktx_read PRIVATE basisu::basisu_encoder + ) endif() create_version_header(src ktx) From a783b02be60d32ad52a6980d27cfd01e7b47dd91 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Sun, 14 Dec 2025 22:50:04 +0200 Subject: [PATCH 04/20] fixed fmt version tag --- external/fmt.cmake | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/external/fmt.cmake b/external/fmt.cmake index b83aef008f..2ce59325f5 100644 --- a/external/fmt.cmake +++ b/external/fmt.cmake @@ -21,7 +21,7 @@ set(FMT_SYSTEM_HEADERS ON) FetchContent_Declare( fmt GIT_REPOSITORY https://github.com/fmtlib/fmt.git - GIT_TAG "v${FMT_VERSION}" + GIT_TAG ${FMT_VERSION} FIND_PACKAGE_ARGS ) From 0317dc2d6eca6c6cefa7c7bb119dd1cd91b0d562 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Mon, 15 Dec 2025 15:05:37 +0200 Subject: [PATCH 05/20] fixed zstd lib install on linux --- external/astcenc.cmake | 2 +- external/zstd.cmake | 2 +- lib/cmake/KtxConfig.cmake | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/external/astcenc.cmake b/external/astcenc.cmake index 10165cfef3..6bc543003a 100644 --- a/external/astcenc.cmake +++ b/external/astcenc.cmake @@ -24,7 +24,6 @@ message(STATUS "ASTC VERSION: ${ASTC_VERSION}") # value of CMAKE_OSX_ARCHITECTURES to decide if a universal build # has been requested. Do not expose the astc-encoder's # ASTCENC_UNIVERSAL_BUILD configuration option. - set(universal_build OFF) if("${CMAKE_SYSTEM_NAME}" STREQUAL "Darwin") # Check CMAKE_OSX_ARCHITECTURES for multiple architectures. @@ -38,6 +37,7 @@ if("${CMAKE_SYSTEM_NAME}" STREQUAL "Darwin") set(ASTCENC_UNIVERSAL_BUILD ${universal_build}) endif() + # If we detect the user is doing a universal build defer to astc-encoder # to pick the architectures. If a universal build has not been requested, # present options to the user to turn off SIMD and, if running on x86_64 diff --git a/external/zstd.cmake b/external/zstd.cmake index 991bb01028..ca7a843198 100644 --- a/external/zstd.cmake +++ b/external/zstd.cmake @@ -30,7 +30,7 @@ FetchContent_Declare( DOWNLOAD_EXTRACT_TIMESTAMP TRUE SOURCE_SUBDIR build/cmake URL "https://github.com/facebook/zstd/releases/download/v${ZSTD_VERSION}/zstd-${ZSTD_VERSION}.tar.gz" - FIND_PACKAGE_ARGS NAMES ZSTD + FIND_PACKAGE_ARGS NAMES ZSTD zstd ) # Populate zstd diff --git a/lib/cmake/KtxConfig.cmake b/lib/cmake/KtxConfig.cmake index 6386ba2f44..bd43c5bdb4 100644 --- a/lib/cmake/KtxConfig.cmake +++ b/lib/cmake/KtxConfig.cmake @@ -1,7 +1,7 @@ # Copyright 2015-2020 The Khronos Group Inc. # SPDX-License-Identifier: Apache-2.0 -# include(CMakeFindDependencyMacro) -# find_dependency() +include(CMakeFindDependencyMacro) +find_dependency(zstd CONFIG) include("${CMAKE_CURRENT_LIST_DIR}/KtxTargets.cmake") From 09a7d7ea21f082e25dc26e0fd4bb89090c1a6bbd Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Mon, 15 Dec 2025 15:51:33 +0200 Subject: [PATCH 06/20] changed astc alias name --- external/astcenc.cmake | 20 +++++++++++--------- lib/CMakeLists.txt | 4 ++-- tests/ktxdiff/CMakeLists.txt | 2 +- tools/ktx/CMakeLists.txt | 2 +- 4 files changed, 15 insertions(+), 13 deletions(-) diff --git a/external/astcenc.cmake b/external/astcenc.cmake index 6bc543003a..8a6daaf15f 100644 --- a/external/astcenc.cmake +++ b/external/astcenc.cmake @@ -4,7 +4,7 @@ include(FetchContent) include(${KTX_ROOT_DIR}/cmake/compiler_query_genexs.cmake) -if (TARGET astc::astcenc) +if (TARGET astcenc::astcenc) message(STATUS "Using prebuilt astc") return() endif() @@ -110,14 +110,16 @@ FetchContent_Declare( # Populate astc FetchContent_MakeAvailable(astcenc) - set(BUILD_SHARED_LIBS ${BUILD_SHARED_LIBS_RESET}) -set_property(TARGET ${ASTCENC_LIB_TARGET} PROPERTY POSITION_INDEPENDENT_CODE ON) -target_compile_definitions( - ${ASTCENC_LIB_TARGET} PRIVATE - $<$,19.40.33811>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> - $<$,17.0.3>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> -) +if (NOT TARGET astcenc::astcenc) + set_property(TARGET ${ASTCENC_LIB_TARGET} PROPERTY POSITION_INDEPENDENT_CODE ON) -add_library(astc::astcenc ALIAS ${ASTCENC_LIB_TARGET}) + target_compile_definitions( + ${ASTCENC_LIB_TARGET} PRIVATE + $<$,19.40.33811>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> + $<$,17.0.3>>:_DISABLE_CONSTEXPR_MUTEX_CONSTRUCTOR> + ) + + add_library(astcenc::astcenc ALIAS ${ASTCENC_LIB_TARGET}) +endif() diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index f0acadce75..3eea66af34 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -475,7 +475,7 @@ macro(common_libktx_settings target enable_write library_type) target_link_libraries( ${target} PRIVATE - $ + $ ) endmacro(common_libktx_settings) @@ -549,7 +549,7 @@ if(LIBKTX_VERSION_READ_ONLY AND NOT (BUILD_SHARED_LIBRARIES OR LIBKTX_VERSION_FU list(APPEND LIBKTX_INSTALL_TARGETS ktx_read) endif() if(NOT BUILD_SHARED_LIBS AND NOT APPLE) - get_target_property(ASTC_LIB_NAME astc::astcenc ALIASED_TARGET) + get_target_property(ASTC_LIB_NAME astcenc::astcenc ALIASED_TARGET) get_target_property(BASISU_LIB_NAME basisu::basisu_encoder ALIASED_TARGET) list(APPEND LIBKTX_INSTALL_TARGETS ${ASTC_LIB_NAME} ${BASISU_LIB_NAME}) diff --git a/tests/ktxdiff/CMakeLists.txt b/tests/ktxdiff/CMakeLists.txt index d5e978702b..af07bfe94b 100644 --- a/tests/ktxdiff/CMakeLists.txt +++ b/tests/ktxdiff/CMakeLists.txt @@ -33,7 +33,7 @@ target_link_libraries( ktxdiff PRIVATE ktx - astc::astcenc + astcenc::astcenc fmt::fmt objUtil ) diff --git a/tools/ktx/CMakeLists.txt b/tools/ktx/CMakeLists.txt index 56d4ba7757..f2890e5f4d 100644 --- a/tools/ktx/CMakeLists.txt +++ b/tools/ktx/CMakeLists.txt @@ -64,7 +64,7 @@ PRIVATE objUtil ktx basisu::basisu_encoder - astc::astcenc + astcenc::astcenc $,Pathcch,> # For PathCchRemoveFileSpec on Windows fmt::fmt cxxopts::cxxopts From d344c3b6922eea3ae88b35181fde69ecf96a4d28 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Mon, 15 Dec 2025 16:55:39 +0200 Subject: [PATCH 07/20] fixed astc install target --- lib/CMakeLists.txt | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index 3eea66af34..16cba11b6f 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -549,10 +549,13 @@ if(LIBKTX_VERSION_READ_ONLY AND NOT (BUILD_SHARED_LIBRARIES OR LIBKTX_VERSION_FU list(APPEND LIBKTX_INSTALL_TARGETS ktx_read) endif() if(NOT BUILD_SHARED_LIBS AND NOT APPLE) - get_target_property(ASTC_LIB_NAME astcenc::astcenc ALIASED_TARGET) get_target_property(BASISU_LIB_NAME basisu::basisu_encoder ALIASED_TARGET) + list(APPEND LIBKTX_INSTALL_TARGETS ${BASISU_LIB_NAME}) - list(APPEND LIBKTX_INSTALL_TARGETS ${ASTC_LIB_NAME} ${BASISU_LIB_NAME}) + get_target_property(ASTC_LIB_NAME astcenc::astcenc ALIASED_TARGET) + if(NOT ${ASTC_LIB_NAME} STREQUAL "ASTC_LIB_NAME-NOTFOUND") + list(APPEND LIBKTX_INSTALL_TARGETS ${ASTC_LIB_NAME}) + endif() endif() if(APPLE OR LINUX) From c3265b5e02f227b6c68aee1ed1f5b920650245b2 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Tue, 16 Dec 2025 15:23:06 +0200 Subject: [PATCH 08/20] more complex library type handling --- external/zstd.cmake | 66 +++++++++++++++++++++++++++++++++------------ 1 file changed, 49 insertions(+), 17 deletions(-) diff --git a/external/zstd.cmake b/external/zstd.cmake index ca7a843198..f0248de52f 100644 --- a/external/zstd.cmake +++ b/external/zstd.cmake @@ -4,12 +4,11 @@ include(FetchContent) if (TARGET zstd::libzstd) - message(STATUS "Using prebuilt zstd") + message(STATUS "(${PROJECT_NAME}): Using configured zstd target") return() endif() set(ZSTD_VERSION 1.5.7) -message(STATUS "ZSTD VERSION: ${ZSTD_VERSION}") # Options # Build static by default @@ -23,24 +22,57 @@ set(ZSTD_BUILD_DEPRECATED OFF) set(ZSTD_BUILD_PROGRAMS OFF) set(ZSTD_BUILD_TESTS OFF) set(ZSTD_BUILD_CONTRIB OFF) +set(CMAKE_POSITION_INDEPENDENT_CODE ON) -# Declare package -FetchContent_Declare( - zstd - DOWNLOAD_EXTRACT_TIMESTAMP TRUE - SOURCE_SUBDIR build/cmake - URL "https://github.com/facebook/zstd/releases/download/v${ZSTD_VERSION}/zstd-${ZSTD_VERSION}.tar.gz" - FIND_PACKAGE_ARGS NAMES ZSTD zstd -) +# On most platforms, static libraries require compilation with -fPIC for shared builds, but zstd is usually supplied without it. +# So we need to build it manually in such cases +# On Windows, this behavior is simplified, so we can try to use system installed library (shared/static) without any fear. +# Or if this is a static library build or was requested to use shared zstd, we can also search for already installed files. +if (NOT ${BUILD_SHARED_LIBS} OR CMAKE_SYSTEM_NAME STREQUAL "Windows" OR ZSTD_BUILD_SHARED) + find_package(zstd CONFIG) +endif() + +if (zstd_FOUND) + message(STATUS "Using system ZSTD ${zstd_VERSION}") +else() + message(STATUS "ZSTD VERSION: ${ZSTD_VERSION}") -# Populate zstd -FetchContent_MakeAvailable(zstd) + # Declare package + FetchContent_Declare( + zstd + DOWNLOAD_EXTRACT_TIMESTAMP TRUE + SOURCE_SUBDIR build/cmake + URL "https://github.com/facebook/zstd/releases/download/v${ZSTD_VERSION}/zstd-${ZSTD_VERSION}.tar.gz" + ) + + # Populate zstd + FetchContent_MakeAvailable(zstd) +endif() if (NOT TARGET zstd::libzstd) - # Create one common target name - if (TARGET libzstd_static) - add_library(zstd::libzstd ALIAS libzstd_static) - elseif(TARGET libzstd_shared) - add_library(zstd::libzstd ALIAS libzstd_shared) + # Normalize different zstd target names to one common alias + set(ZSTD_LOOKUP_NAMES + ZSTD::ZSTD + ) + + if (${ZSTD_BUILD_SHARED}) + list(APPEND ZSTD_LOOKUP_NAMES + zstd::libzstd_shared + libzstd_shared + ) + endif() + + if (${ZSTD_BUILD_STATIC}) + list(APPEND ZSTD_LOOKUP_NAMES + zstd::libzstd_static + libzstd_static + ) endif() + + foreach(LOOKUP_NAME IN LISTS ZSTD_LOOKUP_NAMES) + if (TARGET ${LOOKUP_NAME}) + add_library(zstd::libzstd ALIAS ${LOOKUP_NAME}) + break() + endif() + endforeach() endif() \ No newline at end of file From 928ae83c9d80cd07ea6e568a69802eef405dc281 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Tue, 16 Dec 2025 17:50:20 +0200 Subject: [PATCH 09/20] fixed basisu install target --- lib/CMakeLists.txt | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index 16cba11b6f..8e2a0d6f69 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -474,8 +474,7 @@ macro(common_libktx_settings target enable_write library_type) endif() target_link_libraries( - ${target} PRIVATE - $ + ${target} PRIVATE zstd::libzstd astcenc::astcenc ) endmacro(common_libktx_settings) @@ -548,9 +547,11 @@ endif() if(LIBKTX_VERSION_READ_ONLY AND NOT (BUILD_SHARED_LIBRARIES OR LIBKTX_VERSION_FULL)) list(APPEND LIBKTX_INSTALL_TARGETS ktx_read) endif() -if(NOT BUILD_SHARED_LIBS AND NOT APPLE) +if(NOT BUILD_SHARED_LIBS) get_target_property(BASISU_LIB_NAME basisu::basisu_encoder ALIASED_TARGET) - list(APPEND LIBKTX_INSTALL_TARGETS ${BASISU_LIB_NAME}) + if(NOT ${BASISU_LIB_NAME} STREQUAL "BASISU_LIB_NAME-NOTFOUND") + list(APPEND LIBKTX_INSTALL_TARGETS ${BASISU_LIB_NAME}) + endif() get_target_property(ASTC_LIB_NAME astcenc::astcenc ALIASED_TARGET) if(NOT ${ASTC_LIB_NAME} STREQUAL "ASTC_LIB_NAME-NOTFOUND") From a10a4722adfa45ffe9351598d726c238a816292b Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 00:50:50 +0200 Subject: [PATCH 10/20] dependency files cleanup also fixed dependecies linking for test tools --- external/astcenc.cmake | 6 +++--- external/basisu.cmake | 7 ++++--- external/cxxopts.cmake | 2 +- external/fmt.cmake | 2 +- external/lodepng.cmake | 2 +- tests/ktxdiff/CMakeLists.txt | 1 + tests/transcodetests/CMakeLists.txt | 2 ++ tools/ktx/CMakeLists.txt | 2 ++ 8 files changed, 15 insertions(+), 9 deletions(-) diff --git a/external/astcenc.cmake b/external/astcenc.cmake index 8a6daaf15f..c043ee688e 100644 --- a/external/astcenc.cmake +++ b/external/astcenc.cmake @@ -1,11 +1,11 @@ #################################################### -# ASTC Encoder. +# ASTC Encoder #################################################### include(FetchContent) -include(${KTX_ROOT_DIR}/cmake/compiler_query_genexs.cmake) +include(${CMAKE_CURRENT_LIST_DIR}/../cmake/compiler_query_genexs.cmake) if (TARGET astcenc::astcenc) - message(STATUS "Using prebuilt astc") + message(STATUS "(${PROJECT_NAME}): Using configured ASTC Encoder target") return() endif() diff --git a/external/basisu.cmake b/external/basisu.cmake index 6ec6aa7a79..b14f84a411 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -1,13 +1,14 @@ #################################################### -# Basis Universal Encoder. +# Basis Universal Encoder #################################################### include(FetchContent) if (TARGET basisu::basisu_encoder) - message(STATUS "Using prebuilt basisu") + message(STATUS "(${PROJECT_NAME}): Using configured Basis Universal Encoder target") return() endif() +# TODO: Restore after new basisu release #set(BASISU_VERSION 1_60) #message(STATUS "Basisu VERSION: ${BASISU_VERSION}") @@ -24,7 +25,7 @@ FetchContent_Declare( basisu GIT_REPOSITORY https://github.com/Daniil-SV/basis_universal.git GIT_TAG master #"v${BASISU_VERSION}" - #FIND_PACKAGE_ARGS + FIND_PACKAGE_ARGS ) # Populate basisu diff --git a/external/cxxopts.cmake b/external/cxxopts.cmake index c4a90a48ac..057792b7da 100644 --- a/external/cxxopts.cmake +++ b/external/cxxopts.cmake @@ -4,7 +4,7 @@ include(FetchContent) if (TARGET cxxopts::cxxopts) - message(STATUS "Using prebuilt cxxopts") + message(STATUS "(${PROJECT_NAME}): Using configured cxxopts target") return() endif() diff --git a/external/fmt.cmake b/external/fmt.cmake index 2ce59325f5..153d61836f 100644 --- a/external/fmt.cmake +++ b/external/fmt.cmake @@ -4,7 +4,7 @@ include(FetchContent) if (TARGET fmt::fmt) - message(STATUS "Using prebuilt fmt") + message(STATUS "(${PROJECT_NAME}): Using configured fmt target") return() endif() diff --git a/external/lodepng.cmake b/external/lodepng.cmake index 6f9dab0c24..1b10f091d4 100644 --- a/external/lodepng.cmake +++ b/external/lodepng.cmake @@ -4,7 +4,7 @@ include(FetchContent) if (TARGET lodepng::lodepng) - message(STATUS "Using prebuilt lodepng") + message(STATUS "(${PROJECT_NAME}): Using configured lodepng target") return() endif() diff --git a/tests/ktxdiff/CMakeLists.txt b/tests/ktxdiff/CMakeLists.txt index af07bfe94b..df5acada2a 100644 --- a/tests/ktxdiff/CMakeLists.txt +++ b/tests/ktxdiff/CMakeLists.txt @@ -2,6 +2,7 @@ # Copyright 2022-2023 RasterGrid Kft. # SPDX-License-Identifier: Apache-2.0 +include("../../external/astcenc.cmake") add_executable(ktxdiff ktxdiff_main.cpp diff --git a/tests/transcodetests/CMakeLists.txt b/tests/transcodetests/CMakeLists.txt index 2d98739422..f99f701b53 100644 --- a/tests/transcodetests/CMakeLists.txt +++ b/tests/transcodetests/CMakeLists.txt @@ -1,6 +1,8 @@ # Copyright 2017-2020 The Khronos Group Inc. # SPDX-License-Identifier: Apache-2.0 +include("../../external/basisu.cmake") + add_executable( transcodetests transcodetests.cc ../tests.cmake diff --git a/tools/ktx/CMakeLists.txt b/tools/ktx/CMakeLists.txt index f2890e5f4d..699bcba678 100644 --- a/tools/ktx/CMakeLists.txt +++ b/tools/ktx/CMakeLists.txt @@ -2,6 +2,8 @@ # Copyright 2022-2023 RasterGrid Kft. # SPDX-License-Identifier: Apache-2.0 +include("../../external/basisu.cmake") +include("../../external/astcenc.cmake") add_executable(ktxtools command.cpp From 2ef0ce9301c3979a2d59c728b46ba3aedd24df75 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 01:22:32 +0200 Subject: [PATCH 11/20] added licensing info --- REUSE.toml | 8 +++++++- external/astcenc.cmake | 3 +++ external/basisu.cmake | 3 +++ external/cxxopts.cmake | 3 +++ external/fmt.cmake | 3 +++ external/lodepng.cmake | 3 +++ external/zstd.cmake | 3 +++ 7 files changed, 25 insertions(+), 1 deletion(-) diff --git a/REUSE.toml b/REUSE.toml index d021ca9d06..4b9839fa1f 100644 --- a/REUSE.toml +++ b/REUSE.toml @@ -114,7 +114,7 @@ SPDX-FileCopyrightText = "2017 Sean Barrett" SPDX-License-Identifier = "MIT" [[annotations]] -path = "external/basisu/Source/tinyexr.h" +path = "tools/imageio/external/tinyexr.h" precedence = "aggregate" SPDX-FileCopyrightText = "2014-2019 Syoyo Fujita and many contributors" SPDX-License-Identifier = "BSD-3-Clause" @@ -274,3 +274,9 @@ path = ["vcpkg.json", "vcpkg-configuration.json"] precedence = "aggregate" SPDX-FileCopyrightText = "2024 The Khronos Group Inc." SPDX-License-Identifier = "Apache-2.0" + +[[annotations]] +path = ["tools/imageio/external/jpgd.h", "tools/imageio/external/jpgd.cpp"] +precedence = "aggregate" +SPDX-FileCopyrightText = "1994-2020 Rich Geldreich " +SPDX-License-Identifier = "Apache-2.0" diff --git a/external/astcenc.cmake b/external/astcenc.cmake index c043ee688e..8363cb33dd 100644 --- a/external/astcenc.cmake +++ b/external/astcenc.cmake @@ -1,3 +1,6 @@ +# Copyright 2025 Danylo Sivachenko (@Daniil-SV). +# SPDX-License-Identifier: Apache-2.0 + #################################################### # ASTC Encoder #################################################### diff --git a/external/basisu.cmake b/external/basisu.cmake index b14f84a411..99edf58304 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -1,3 +1,6 @@ +# Copyright 2025 Danylo Sivachenko (@Daniil-SV). +# SPDX-License-Identifier: Apache-2.0 + #################################################### # Basis Universal Encoder #################################################### diff --git a/external/cxxopts.cmake b/external/cxxopts.cmake index 057792b7da..8e6b5f619e 100644 --- a/external/cxxopts.cmake +++ b/external/cxxopts.cmake @@ -1,3 +1,6 @@ +# Copyright 2025 Danylo Sivachenko (@Daniil-SV). +# SPDX-License-Identifier: Apache-2.0 + #################################################### # cxxopts #################################################### diff --git a/external/fmt.cmake b/external/fmt.cmake index 153d61836f..36c9978a90 100644 --- a/external/fmt.cmake +++ b/external/fmt.cmake @@ -1,3 +1,6 @@ +# Copyright 2025 Danylo Sivachenko (@Daniil-SV). +# SPDX-License-Identifier: Apache-2.0 + #################################################### # fmt #################################################### diff --git a/external/lodepng.cmake b/external/lodepng.cmake index 1b10f091d4..73173b60a9 100644 --- a/external/lodepng.cmake +++ b/external/lodepng.cmake @@ -1,3 +1,6 @@ +# Copyright 2025 Danylo Sivachenko (@Daniil-SV). +# SPDX-License-Identifier: Apache-2.0 + #################################################### # lodepng #################################################### diff --git a/external/zstd.cmake b/external/zstd.cmake index f0248de52f..747665d50a 100644 --- a/external/zstd.cmake +++ b/external/zstd.cmake @@ -1,3 +1,6 @@ +# Copyright 2025 Danylo Sivachenko (@Daniil-SV). +# SPDX-License-Identifier: Apache-2.0 + #################################################### # Zstandart #################################################### From 0cd072ff814874c8035871301ffe29fec49da563 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 01:26:02 +0200 Subject: [PATCH 12/20] disabled werror for ci --- .github/workflows/android.yml | 2 +- .github/workflows/check-libktx-only.yml | 2 +- .github/workflows/check-mkvk.yml | 2 +- .github/workflows/mingw.yml | 2 +- .github/workflows/windows.yml | 2 +- 5 files changed, 5 insertions(+), 5 deletions(-) diff --git a/.github/workflows/android.yml b/.github/workflows/android.yml index e1c317ce98..0d214493ea 100644 --- a/.github/workflows/android.yml +++ b/.github/workflows/android.yml @@ -56,7 +56,7 @@ jobs: env: GIT_LFS_SKIP_SMUDGE: 1 - WERROR: ON + WERROR: OFF steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/check-libktx-only.yml b/.github/workflows/check-libktx-only.yml index a724eed013..27d8028c53 100644 --- a/.github/workflows/check-libktx-only.yml +++ b/.github/workflows/check-libktx-only.yml @@ -88,7 +88,7 @@ jobs: # In theory it should not matter, but this image is chosen to exercise # some of the functions in one of the library dependencies. KTX2_TEST_FILE: "tests/testimages/kodim17_basis.ktx2" - WERROR: ON + WERROR: OFF steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/check-mkvk.yml b/.github/workflows/check-mkvk.yml index 808b8d1666..2af027f8d0 100644 --- a/.github/workflows/check-mkvk.yml +++ b/.github/workflows/check-mkvk.yml @@ -56,7 +56,7 @@ jobs: env: GIT_LFS_SKIP_SMUDGE: 1 BUILD_DIR: "build" - WERROR: ON + WERROR: OFF steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/mingw.yml b/.github/workflows/mingw.yml index 1f2d983dbb..5d39d796f0 100644 --- a/.github/workflows/mingw.yml +++ b/.github/workflows/mingw.yml @@ -56,7 +56,7 @@ jobs: # for texturetests etc., loadtests and legacy tool tests. They will be # pulled later as necessary GIT_LFS_SKIP_SMUDGE: 1 - WERROR: ON + WERROR: OFF steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/windows.yml b/.github/workflows/windows.yml index c9c5639c2a..40fee1a442 100644 --- a/.github/workflows/windows.yml +++ b/.github/workflows/windows.yml @@ -140,7 +140,7 @@ jobs: SUPPORT_OPENCL: ${{ matrix.options.opencl }} SUPPORT_SSE: ${{ matrix.options.sse }} PY_USE_VENV: ${{ matrix.options.py.use_venv && matrix.options.py_use_venv || 'OFF' }} - WERROR: ON + WERROR: OFF VCPKG_INSTALL_OPTIONS: ${{ matrix.vcpkg_install_options && matrix.vcpkg_install_options || '' }} # This is set to avoid the compiler running out of heap space From 54265010b7fa6f883a4780fefca047fa0e41e563 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 02:52:40 +0200 Subject: [PATCH 13/20] added the rest of basisu options --- external/basisu.cmake | 2 ++ 1 file changed, 2 insertions(+) diff --git a/external/basisu.cmake b/external/basisu.cmake index 99edf58304..29f8724055 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -18,6 +18,8 @@ endif() # Options set(TOOL FALSE) set(EXAMPLES FALSE) +set(SSE ${BASISU_SSE}) +set(OPENCL ${BASISU_OPENCL}) if(NOT ${CPU_ARCHITECTURE} STREQUAL "x86_64") # Basisu sets this TRUE if MSVC is TRUE. set(BASISU_SSE FALSE) From 3dbffee8e9ccf5eebd35aaed89fe2fb9cebad5e7 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 14:24:20 +0200 Subject: [PATCH 14/20] changed astc git url to raw archive url --- external/astcenc.cmake | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/external/astcenc.cmake b/external/astcenc.cmake index 8363cb33dd..f0f27f080a 100644 --- a/external/astcenc.cmake +++ b/external/astcenc.cmake @@ -106,8 +106,8 @@ set(BUILD_SHARED_LIBS OFF) # Declare package FetchContent_Declare( astcenc - GIT_REPOSITORY https://github.com/ARM-software/astc-encoder - GIT_TAG ${ASTC_VERSION} + DOWNLOAD_EXTRACT_TIMESTAMP TRUE + URL "https://github.com/ARM-software/astc-encoder/archive/refs/tags/${ASTC_VERSION}.tar.gz" FIND_PACKAGE_ARGS ) From 14b9d6f3dc3726e74ed2c6cabc7ed7069f97c604 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 14:24:45 +0200 Subject: [PATCH 15/20] disabled werror for macos ci --- .github/workflows/macos.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/macos.yml b/.github/workflows/macos.yml index e42fd6dfcf..bce10a32b3 100644 --- a/.github/workflows/macos.yml +++ b/.github/workflows/macos.yml @@ -137,7 +137,7 @@ jobs: WASM_BUILD: ${{ matrix.options.wasm && matrix.options.wasm || 'OFF' }} # Virtual env (ON) is needed because pre-installed python is "externally managed." PY_USE_VENV: ${{ matrix.options.py.use_venv && matrix.options.py_use_venv || 'ON' }} - WERROR: ${{ matrix.options.werror && matrix.options.werror || 'ON' }} + WERROR: ${{ matrix.options.werror && matrix.options.werror || 'OFF' }} # CC Handled by job step # CXX Handled by job step From 3b25f388d3f51282baaee04e037cc69843ee6b5e Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 15:56:36 +0200 Subject: [PATCH 16/20] fixed werror behavior on some dependencies --- .github/workflows/android.yml | 2 +- .github/workflows/check-libktx-only.yml | 2 +- .github/workflows/check-mkvk.yml | 2 +- .github/workflows/macos.yml | 2 +- .github/workflows/mingw.yml | 2 +- .github/workflows/windows.yml | 2 +- external/basisu.cmake | 16 +++++++++++++--- 7 files changed, 19 insertions(+), 9 deletions(-) diff --git a/.github/workflows/android.yml b/.github/workflows/android.yml index 0d214493ea..e1c317ce98 100644 --- a/.github/workflows/android.yml +++ b/.github/workflows/android.yml @@ -56,7 +56,7 @@ jobs: env: GIT_LFS_SKIP_SMUDGE: 1 - WERROR: OFF + WERROR: ON steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/check-libktx-only.yml b/.github/workflows/check-libktx-only.yml index 27d8028c53..a724eed013 100644 --- a/.github/workflows/check-libktx-only.yml +++ b/.github/workflows/check-libktx-only.yml @@ -88,7 +88,7 @@ jobs: # In theory it should not matter, but this image is chosen to exercise # some of the functions in one of the library dependencies. KTX2_TEST_FILE: "tests/testimages/kodim17_basis.ktx2" - WERROR: OFF + WERROR: ON steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/check-mkvk.yml b/.github/workflows/check-mkvk.yml index 2af027f8d0..808b8d1666 100644 --- a/.github/workflows/check-mkvk.yml +++ b/.github/workflows/check-mkvk.yml @@ -56,7 +56,7 @@ jobs: env: GIT_LFS_SKIP_SMUDGE: 1 BUILD_DIR: "build" - WERROR: OFF + WERROR: ON steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/macos.yml b/.github/workflows/macos.yml index bce10a32b3..e42fd6dfcf 100644 --- a/.github/workflows/macos.yml +++ b/.github/workflows/macos.yml @@ -137,7 +137,7 @@ jobs: WASM_BUILD: ${{ matrix.options.wasm && matrix.options.wasm || 'OFF' }} # Virtual env (ON) is needed because pre-installed python is "externally managed." PY_USE_VENV: ${{ matrix.options.py.use_venv && matrix.options.py_use_venv || 'ON' }} - WERROR: ${{ matrix.options.werror && matrix.options.werror || 'OFF' }} + WERROR: ${{ matrix.options.werror && matrix.options.werror || 'ON' }} # CC Handled by job step # CXX Handled by job step diff --git a/.github/workflows/mingw.yml b/.github/workflows/mingw.yml index 5d39d796f0..1f2d983dbb 100644 --- a/.github/workflows/mingw.yml +++ b/.github/workflows/mingw.yml @@ -56,7 +56,7 @@ jobs: # for texturetests etc., loadtests and legacy tool tests. They will be # pulled later as necessary GIT_LFS_SKIP_SMUDGE: 1 - WERROR: OFF + WERROR: ON steps: - uses: actions/checkout@v4 diff --git a/.github/workflows/windows.yml b/.github/workflows/windows.yml index 40fee1a442..c9c5639c2a 100644 --- a/.github/workflows/windows.yml +++ b/.github/workflows/windows.yml @@ -140,7 +140,7 @@ jobs: SUPPORT_OPENCL: ${{ matrix.options.opencl }} SUPPORT_SSE: ${{ matrix.options.sse }} PY_USE_VENV: ${{ matrix.options.py.use_venv && matrix.options.py_use_venv || 'OFF' }} - WERROR: OFF + WERROR: ON VCPKG_INSTALL_OPTIONS: ${{ matrix.vcpkg_install_options && matrix.vcpkg_install_options || '' }} # This is set to avoid the compiler running out of heap space diff --git a/external/basisu.cmake b/external/basisu.cmake index 29f8724055..9893ec12ab 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -18,12 +18,12 @@ endif() # Options set(TOOL FALSE) set(EXAMPLES FALSE) -set(SSE ${BASISU_SSE}) -set(OPENCL ${BASISU_OPENCL}) if(NOT ${CPU_ARCHITECTURE} STREQUAL "x86_64") # Basisu sets this TRUE if MSVC is TRUE. set(BASISU_SSE FALSE) endif() +set(SSE ${BASISU_SSE}) +set(OPENCL ${BASISU_OPENCL}) # Declare package FetchContent_Declare( @@ -34,7 +34,17 @@ FetchContent_Declare( ) # Populate basisu -FetchContent_MakeAvailable(basisu) +# We need to explicitly populate basisu and add it as a subdirectory with EXCLUDE_FROM_ALL +# to avoid problems with KTX_WERROR in ci +FetchContent_GetProperties(basisu) +if (NOT basisu_POPULATED) + FetchContent_Populate(basisu) + add_subdirectory( + ${basisu_SOURCE_DIR} + ${basisu_BINARY_DIR} + EXCLUDE_FROM_ALL + ) +endif() if (NOT TARGET basisu::basisu_encoder) add_library(basisu::basisu_encoder ALIAS basisu_encoder) From db895da5812ab8e9a53f708be7e88607c07a4e69 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 17 Dec 2025 21:55:50 +0200 Subject: [PATCH 17/20] a try to fix WASM CI --- CMakeLists.txt | 4 ++-- external/astcenc.cmake | 4 ++-- external/basisu.cmake | 12 +----------- interface/js_binding/transcoder_wrapper.cpp | 2 +- scripts/build_wasm_docker.sh | 14 ++++++++++++-- 5 files changed, 18 insertions(+), 18 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index 1ac3e3822a..395f852cc0 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -173,14 +173,14 @@ endif() # Global compile & link options including optimization flags if(MSVC) - add_compile_options( /W4;$<$:/WX> ) + #add_compile_options( /W4;$<$:/WX> ) add_compile_options( $,/Gz,/O2> ) # Enable UTF-8 support add_compile_options( $<$:/utf-8> ) add_compile_options( $<$:/utf-8> ) elseif(${CMAKE_CXX_COMPILER_ID} STREQUAL "GNU" OR ${CMAKE_CXX_COMPILER_ID} MATCHES "Clang") - add_compile_options( -Wall -Wextra $<$:-Werror>) + #add_compile_options( -Wall -Wextra $<$:-Werror>) add_compile_options( $,-O0$-g,-O3> ) if(EMSCRIPTEN) add_link_options( $,-gsource-map,-O3> ) diff --git a/external/astcenc.cmake b/external/astcenc.cmake index f0f27f080a..8363cb33dd 100644 --- a/external/astcenc.cmake +++ b/external/astcenc.cmake @@ -106,8 +106,8 @@ set(BUILD_SHARED_LIBS OFF) # Declare package FetchContent_Declare( astcenc - DOWNLOAD_EXTRACT_TIMESTAMP TRUE - URL "https://github.com/ARM-software/astc-encoder/archive/refs/tags/${ASTC_VERSION}.tar.gz" + GIT_REPOSITORY https://github.com/ARM-software/astc-encoder + GIT_TAG ${ASTC_VERSION} FIND_PACKAGE_ARGS ) diff --git a/external/basisu.cmake b/external/basisu.cmake index 9893ec12ab..a085274d83 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -34,17 +34,7 @@ FetchContent_Declare( ) # Populate basisu -# We need to explicitly populate basisu and add it as a subdirectory with EXCLUDE_FROM_ALL -# to avoid problems with KTX_WERROR in ci -FetchContent_GetProperties(basisu) -if (NOT basisu_POPULATED) - FetchContent_Populate(basisu) - add_subdirectory( - ${basisu_SOURCE_DIR} - ${basisu_BINARY_DIR} - EXCLUDE_FROM_ALL - ) -endif() +FetchContent_MakeAvailable(basisu) if (NOT TARGET basisu::basisu_encoder) add_library(basisu::basisu_encoder ALIAS basisu_encoder) diff --git a/interface/js_binding/transcoder_wrapper.cpp b/interface/js_binding/transcoder_wrapper.cpp index b54b8aaab7..b7caa384dd 100644 --- a/interface/js_binding/transcoder_wrapper.cpp +++ b/interface/js_binding/transcoder_wrapper.cpp @@ -12,7 +12,7 @@ */ #include -#include "transcoder/basisu_transcoder.h" +#include "basisu/transcoder/basisu_transcoder.h" using namespace emscripten; using namespace basist; diff --git a/scripts/build_wasm_docker.sh b/scripts/build_wasm_docker.sh index 2b076a07ba..81cdfca7b1 100755 --- a/scripts/build_wasm_docker.sh +++ b/scripts/build_wasm_docker.sh @@ -136,7 +136,17 @@ if [ -n "$FEATURE_LOADTESTS" ]; then docker exec emscripten sh -c "embuilder build sdl3" fi -docker exec emscripten sh -c "emcmake cmake -B$BUILD_DIR . \ +# The Docker container's integrated cmake is acting very strangely. +# It generates incorrect files for fetching dependencies via FetchContent, causing the build and CI to fail. +# As a workaround, we can download the latest cmake build from official website +# and install it during Docker container configuration, then use it to configure our package. +docker exec -u root emscripten sh -c "mkdir -p /opt/cmake && \ + wget https://github.com/Kitware/CMake/releases/download/v3.31.10/cmake-3.31.10-linux-x86_64.sh \ + -O /tmp/cmake.sh -q && \ + chmod +x /tmp/cmake.sh && \ + sh /tmp/cmake.sh --skip-license --prefix=/opt/cmake" + +docker exec emscripten sh -c "emcmake /opt/cmake/bin/cmake -B$BUILD_DIR . \ -D CMAKE_BUILD_TYPE=$CONFIGURATION \ -D KTX_FEATURE_DOC=$FEATURE_DOC \ -D KTX_FEATURE_JNI=$FEATURE_JNI \ @@ -144,7 +154,7 @@ docker exec emscripten sh -c "emcmake cmake -B$BUILD_DIR . \ -D KTX_FEATURE_TESTS=$FEATURE_TESTS \ -D KTX_FEATURE_TOOLS=$FEATURE_TOOLS \ -D KTX_WERROR=$WERROR \ - && cmake --build $BUILD_DIR $targets $verbose_make" + && /opt/cmake/bin/cmake --build $BUILD_DIR $targets $verbose_make" if [ "$PACKAGE" = "YES" ]; then echo "Pack KTX-Software (Web $CONFIGURATION)" From d01fa7f5c08a82b722e92bbdc1e6acae92b4ce74 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Tue, 23 Dec 2025 00:04:37 +0200 Subject: [PATCH 18/20] added back library optimization updated basisu source tag improved zstd module behavior --- external/basisu.cmake | 8 +++----- external/zstd.cmake | 31 +++++++++++++++++++++++++------ lib/CMakeLists.txt | 36 ++++++++++++++++++++++++++---------- 3 files changed, 54 insertions(+), 21 deletions(-) diff --git a/external/basisu.cmake b/external/basisu.cmake index a085274d83..30433335bd 100644 --- a/external/basisu.cmake +++ b/external/basisu.cmake @@ -16,20 +16,18 @@ endif() #message(STATUS "Basisu VERSION: ${BASISU_VERSION}") # Options -set(TOOL FALSE) -set(EXAMPLES FALSE) +set(BASISU_TOOL FALSE) +set(BASISU_EXAMPLES FALSE) if(NOT ${CPU_ARCHITECTURE} STREQUAL "x86_64") # Basisu sets this TRUE if MSVC is TRUE. set(BASISU_SSE FALSE) endif() -set(SSE ${BASISU_SSE}) -set(OPENCL ${BASISU_OPENCL}) # Declare package FetchContent_Declare( basisu GIT_REPOSITORY https://github.com/Daniil-SV/basis_universal.git - GIT_TAG master #"v${BASISU_VERSION}" + GIT_TAG cmake_fixes #"v${BASISU_VERSION}" FIND_PACKAGE_ARGS ) diff --git a/external/zstd.cmake b/external/zstd.cmake index 747665d50a..672d168777 100644 --- a/external/zstd.cmake +++ b/external/zstd.cmake @@ -27,12 +27,15 @@ set(ZSTD_BUILD_TESTS OFF) set(ZSTD_BUILD_CONTRIB OFF) set(CMAKE_POSITION_INDEPENDENT_CODE ON) -# On most platforms, static libraries require compilation with -fPIC for shared builds, but zstd is usually supplied without it. -# So we need to build it manually in such cases -# On Windows, this behavior is simplified, so we can try to use system installed library (shared/static) without any fear. -# Or if this is a static library build or was requested to use shared zstd, we can also search for already installed files. -if (NOT ${BUILD_SHARED_LIBS} OR CMAKE_SYSTEM_NAME STREQUAL "Windows" OR ZSTD_BUILD_SHARED) - find_package(zstd CONFIG) +# Trying to mimic FetchContent`s find_package behavior, but with our custom rules +if (NOT FETCHCONTENT_TRY_FIND_PACKAGE_MODE STREQUAL "NEVER") + # On most platforms, static libraries require compilation with -fPIC for shared builds, but zstd is usually supplied without it. + # So we need to build it manually in such cases + # On Windows, this behavior is simplified, so we can try to use system installed library (shared/static) without any fear. + # Or if this is a static library build or was requested to use shared zstd, we can also search for already installed files. + if (NOT ${BUILD_SHARED_LIBS} OR CMAKE_SYSTEM_NAME STREQUAL "Windows" OR ZSTD_BUILD_SHARED) + find_package(zstd CONFIG) + endif() endif() if (zstd_FOUND) @@ -75,7 +78,23 @@ if (NOT TARGET zstd::libzstd) foreach(LOOKUP_NAME IN LISTS ZSTD_LOOKUP_NAMES) if (TARGET ${LOOKUP_NAME}) add_library(zstd::libzstd ALIAS ${LOOKUP_NAME}) + + # Disable unused function warning + if (MSVC) + target_compile_options( + ${LOOKUP_NAME} PRIVATE + /wd4505 + $<$: /clang:-Wno-unused-function> + ) + else() + target_compile_options( + ${LOOKUP_NAME} PRIVATE + $<$: -Wno-unused-function> + ) + endif() break() endif() endforeach() + + endif() \ No newline at end of file diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt index 8e2a0d6f69..24fdf1853a 100644 --- a/lib/CMakeLists.txt +++ b/lib/CMakeLists.txt @@ -209,6 +209,29 @@ add_compile_options( ${fp_options} ) include(${KTX_ROOT_DIR}/cmake/compiler_query_genexs.cmake) +# Make a single static library to simplify linking. +macro(add_lib_dependencies target lib) + # still need to add include directories and make sure that everything will be ok with install targets + target_link_libraries(${target} PRIVATE ${lib}) + + # getting original target name + set(library_name ) + get_target_property(aliased_name ${lib} ALIASED_TARGET) + if (aliased_name) + set(library_name ${aliased_name}) + else() + set(library_name ${lib}) + endif() + + # make sure that dependency target is actual target + # there's no sense to do something with imported targets + get_target_property(is_imported ${library_name} IMPORTED) + if(NOT BUILD_SHARED_LIBS AND NOT is_imported) + target_sources(${target} PRIVATE $) + target_include_directories(${target} PRIVATE $) + endif() +endmacro(add_lib_dependencies) + # Macro for common lib settings macro(common_libktx_settings target enable_write library_type) @@ -473,12 +496,11 @@ macro(common_libktx_settings target enable_write library_type) ) endif() - target_link_libraries( - ${target} PRIVATE zstd::libzstd astcenc::astcenc - ) + add_lib_dependencies(${target} zstd::libzstd) + add_lib_dependencies(${target} astcenc::astcenc) + add_lib_dependencies(${target} basisu::basisu_encoder) endmacro(common_libktx_settings) - ################################################### # ktx libraries ################################################### @@ -491,9 +513,6 @@ if(LIBKTX_VERSION_FULL) src/writer2.c ) common_libktx_settings(ktx 1 ${LIB_TYPE}) - target_link_libraries( - ktx PRIVATE basisu::basisu_encoder - ) endif() # Read-only library @@ -502,9 +521,6 @@ if(LIBKTX_VERSION_READ_ONLY) ${LIBKTX_MAIN_SRC} ) common_libktx_settings(ktx_read 0 ${LIB_TYPE}) - target_link_libraries( - ktx_read PRIVATE basisu::basisu_encoder - ) endif() create_version_header(src ktx) From a2070cbf4233fa36e2c7e5a982eb4b8e4ba29846 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 24 Dec 2025 15:06:04 +0200 Subject: [PATCH 19/20] revert for disabled global werror flags --- CMakeLists.txt | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index 395f852cc0..1ac3e3822a 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -173,14 +173,14 @@ endif() # Global compile & link options including optimization flags if(MSVC) - #add_compile_options( /W4;$<$:/WX> ) + add_compile_options( /W4;$<$:/WX> ) add_compile_options( $,/Gz,/O2> ) # Enable UTF-8 support add_compile_options( $<$:/utf-8> ) add_compile_options( $<$:/utf-8> ) elseif(${CMAKE_CXX_COMPILER_ID} STREQUAL "GNU" OR ${CMAKE_CXX_COMPILER_ID} MATCHES "Clang") - #add_compile_options( -Wall -Wextra $<$:-Werror>) + add_compile_options( -Wall -Wextra $<$:-Werror>) add_compile_options( $,-O0$-g,-O3> ) if(EMSCRIPTEN) add_link_options( $,-gsource-map,-O3> ) From 23710975e779b0afd6c29b148d3ec0d76e6f3ed9 Mon Sep 17 00:00:00 2001 From: Daniil-SV Date: Wed, 24 Dec 2025 15:19:09 +0200 Subject: [PATCH 20/20] fixed target_compile_options with wrong target call in zstd module --- external/zstd.cmake | 26 +++++++++++++++----------- 1 file changed, 15 insertions(+), 11 deletions(-) diff --git a/external/zstd.cmake b/external/zstd.cmake index 672d168777..e2436a5982 100644 --- a/external/zstd.cmake +++ b/external/zstd.cmake @@ -80,18 +80,22 @@ if (NOT TARGET zstd::libzstd) add_library(zstd::libzstd ALIAS ${LOOKUP_NAME}) # Disable unused function warning - if (MSVC) - target_compile_options( - ${LOOKUP_NAME} PRIVATE - /wd4505 - $<$: /clang:-Wno-unused-function> - ) - else() - target_compile_options( - ${LOOKUP_NAME} PRIVATE - $<$: -Wno-unused-function> - ) + get_target_property(is_imported ${LOOKUP_NAME} IMPORTED) + if(NOT is_imported) + if (MSVC) + target_compile_options( + ${LOOKUP_NAME} PRIVATE + /wd4505 + $<$: /clang:-Wno-unused-function> + ) + else() + target_compile_options( + ${LOOKUP_NAME} PRIVATE + $<$: -Wno-unused-function> + ) + endif() endif() + break() endif() endforeach()