Tries to download libargparse-populate during build.
On 2025/03/20 20:27, Brad Smith wrote: > Here is an update to libavif 1.2.1. > > > ## [1.2.1] - 2025-03-17 > > ### Added since 1.2.0 > > * Add support for outputting all frames of an image sequence in `avifdec`. > `avifdec --index all sequence.avif out.png` creates files named > `out-xxxxxxxxxx.png` where xxxxxxxxxx are the zero-padded frame indices. > > ### Changed since 1.2.0 > > * Fix local libargparse dependency patch step on macOS 10.15 and earlier. > * Patch local libyuv dependency for compatibility with gcc 10. > * Use stricter C99 syntax to avoid related compilation issues. > > ## [1.2.0] - 2025-02-25 > > ### Added since 1.1.1 > * Turn on the gain map API. Remove the AVIF_ENABLE_EXPERIMENTAL_GAIN_MAP CMake > flag. > * Allow YCgCo_Re and YCgCo_Ro encoding/decoding and update the enum values to > the latest CICP specification. Remove the AVIF_ENABLE_EXPERIMENTAL_YCGCO_R > CMake flag. > * Add the properties and numProperties fields to avifImage. They are filled by > the avifDecoder instance with the properties unrecognized by libavif. They > are > written by the avifEncoder. > * Add avif(Un)SignedFraction structs and avifDoubleTo(Un)SignedFraction > utility functions. > * Add 'avifgainmaputil' command line tool to installed apps. > * Add avifCropRectRequiresUpsampling(). > * Add experimental support for PixelInformationProperty syntax from HEIF 3rd > Ed. > Amd2 behind the compilation flag AVIF_ENABLE_EXPERIMENTAL_EXTENDED_PIXI. > * Add experimental Sample Transform recipe > BIT_DEPTH_EXTENSION_12B_8B_OVERLAP_4B. > > ### Changed since 1.1.1 > * avifenc: Allow large images to be encoded. > * Fix empty CMAKE_CXX_FLAGS_RELEASE if -DAVIF_CODEC_AOM=LOCAL > -DAVIF_LIBYUV=OFF > is specified. https://github.com/AOMediaCodec/libavif/issues/2365. > * Rename AVIF_ENABLE_EXPERIMENTAL_METAV1 to AVIF_ENABLE_EXPERIMENTAL_MINI and > update the experimental reduced header feature to the latest specification > draft. Rename AVIF_HEADER_REDUCED to AVIF_HEADER_MINI. > * Update the experimental Sample Transform feature behind the > AVIF_ENABLE_EXPERIMENTAL_SAMPLE_TRANSFORM CMake flag to the latest > specification draft. > * Ignore gain maps with unsupported metadata. Handle gain maps with > writer_version > 0 correctly. > * Simplify gain map API: remove the enableParsingGainMapMetadata setting, now > gain map metadata is always parsed if present and if this feature is > compiled > in. Replace enableDecodingGainMap and ignoreColorAndAlpha with a bit field > to > choose image content to decode. Remove gainMapPresent: users can check if > decoder->image->gainMap != NULL instead. Remove avifGainMapMetadata and > avifGainMapMetadataDouble structs. > * Write an empty HandlerBox name field instead of "libavif" (saves 7 bytes). > * Check for FileTypeBox precedence in avifParse(). > * Do not write an alternative group with the same ID as an item. > * Update parseAV2SequenceHeader() and avm.cmd: research-v9.0.0 > * Remove AVIF_ENABLE_GTEST CMake option. It's now implied by > AVIF_GTEST=LOCAL/SYSTEM. > * Deprecate `avifEncoder`'s `minQuantizer`, `maxQuantizer`, > `minQuantizerAlpha`, > and `maxQuantizerAlpha` fields. `quality` and `qualityAlpha`\xc2\xa0should > be used > instead. Deprecate\xc2\xa0`avifenc`'s `--min`, `--max`, `--minalpha` and > `--maxalpha` > flags. `-q` or `--qcolor` and `--qalpha`\xc2\xa0should be used instead. > * For dependencies, the deprecated way of setting AVIF_LOCAL_* to ON is > removed. Dependency options can now only be set to OFF/LOCAL/SYSTEM. > * Change the default quality for alpha to be the same as the quality for > color. > * Allow decoding subsampled images with odd Clean Aperture dimensions or > offsets. > * Deprecate avifCropRectConvertCleanApertureBox() and > avifCleanApertureBoxConvertCropRect(). Replace them with > avifCropRectFromCleanApertureBox() and avifCleanApertureBoxFromCropRect(). > * Write descriptive properties before transformative properties. > * Reject non-essential transformative properties. > * Treat avifenc --stdin as a regular positional file path argument. > * Update man pages based on avifenc/dec's --help message. > * android_jni: Support 16kb page size > * android_jni: Set threads to 2 instead of CPU count > * Fix overflows when dealing with alpha during YUV/RGB conversions and in > avifRGBImageAllocatePixels(). > * Make avifEncoder.headerFormat a flag combination for future features. > * Rename AVIF_HEADER_FULL to AVIF_HEADER_DEFAULT. Deprecate AVIF_HEADER_FULL. > * Fix decoding image sequences with non video tracks (such as audio or > subtitles). > * Fix type checking of auxiliary tracks: previously any auxiliary track was > assumed to be alpha, even if it was a different type. If the aux type is > absent, > it is assumed to be alpha. > > > Index: Makefile > =================================================================== > RCS file: /cvs/ports/graphics/libavif/Makefile,v > retrieving revision 1.24 > diff -u -p -u -p -r1.24 Makefile > --- Makefile 30 Aug 2024 07:48:53 -0000 1.24 > +++ Makefile 20 Mar 2025 01:58:40 -0000 > @@ -1,11 +1,11 @@ > COMMENT= library for encoding and decoding .avif files > > -DIST_TUPLE+= github AOMediaCodec libavif v1.1.1 . > +DIST_TUPLE+= github AOMediaCodec libavif v1.2.1 . > DIST_TUPLE+= github lemenkov libyuv \ > - a6a2ec654b1be1166b376476a7555c89eca0c275 ext/libyuv > + ccdf870348764e4b77fa3b56accb2a896a901bad ext/libyuv > CATEGORIES= graphics > > -SHARED_LIBS= avif 7.1 > +SHARED_LIBS= avif 8.0 > > HOMEPAGE= https://github.com/AOMediaCodec/libavif > > Index: distinfo > =================================================================== > RCS file: /cvs/ports/graphics/libavif/distinfo,v > retrieving revision 1.18 > diff -u -p -u -p -r1.18 distinfo > --- distinfo 30 Aug 2024 07:48:53 -0000 1.18 > +++ distinfo 20 Mar 2025 01:58:40 -0000 > @@ -1,4 +1,4 @@ > -SHA256 (AOMediaCodec-libavif-v1.1.1.tar.gz) = > kUZi4WJF4GLtc/kBEvu0VIJBMAhDp3ctjUQbtoWd5Fs= > -SHA256 (lemenkov-libyuv-a6a2ec654b1be1166b376476a7555c89eca0c275.tar.gz) = > +5dXKi3B2Ok3bmGzPrSHCMzPs0Egb8m02atl9Asyb7M= > -SIZE (AOMediaCodec-libavif-v1.1.1.tar.gz) = 13644189 > -SIZE (lemenkov-libyuv-a6a2ec654b1be1166b376476a7555c89eca0c275.tar.gz) = > 570482 > +SHA256 (AOMediaCodec-libavif-v1.2.1.tar.gz) = > nIWcfBLMsPQHURv+MD5qckf19nOPVIUmYsbfgEja3fQ= > +SHA256 (lemenkov-libyuv-ccdf870348764e4b77fa3b56accb2a896a901bad.tar.gz) = > F0H+XOzsy+15Q/nMzwJgHrQwtzF/jUTMaFY5XEaqyuw= > +SIZE (AOMediaCodec-libavif-v1.2.1.tar.gz) = 13942041 > +SIZE (lemenkov-libyuv-ccdf870348764e4b77fa3b56accb2a896a901bad.tar.gz) = > 614050 > Index: patches/patch-ext_libyuv_include_libyuv_cpu_id_h > =================================================================== > RCS file: patches/patch-ext_libyuv_include_libyuv_cpu_id_h > diff -N patches/patch-ext_libyuv_include_libyuv_cpu_id_h > --- /dev/null 1 Jan 1970 00:00:00 -0000 > +++ patches/patch-ext_libyuv_include_libyuv_cpu_id_h 20 Mar 2025 01:58:40 > -0000 > @@ -0,0 +1,12 @@ > +Index: ext/libyuv/include/libyuv/cpu_id.h > +--- ext/libyuv/include/libyuv/cpu_id.h.orig > ++++ ext/libyuv/include/libyuv/cpu_id.h > +@@ -94,7 +94,7 @@ int MipsCpuCaps(const char* cpuinfo_name); > + LIBYUV_API > + int RiscvCpuCaps(const char* cpuinfo_name); > + > +-#ifdef __linux__ > ++#if defined(__linux__) || defined(__FreeBSD__) || defined(__OpenBSD__) > + // On Linux, parse AArch64 features from getauxval(AT_HWCAP{,2}). > + LIBYUV_API > + int AArch64CpuCaps(unsigned long hwcap, unsigned long hwcap2); > Index: patches/patch-ext_libyuv_source_cpu_id_cc > =================================================================== > RCS file: patches/patch-ext_libyuv_source_cpu_id_cc > diff -N patches/patch-ext_libyuv_source_cpu_id_cc > --- /dev/null 1 Jan 1970 00:00:00 -0000 > +++ patches/patch-ext_libyuv_source_cpu_id_cc 20 Mar 2025 01:58:40 -0000 > @@ -0,0 +1,44 @@ > +Index: ext/libyuv/source/cpu_id.cc > +--- ext/libyuv/source/cpu_id.cc.orig > ++++ ext/libyuv/source/cpu_id.cc > +@@ -23,7 +23,8 @@ > + #include <stdio.h> // For fopen() > + #include <string.h> > + > +-#if defined(__linux__) && defined(__aarch64__) > ++#if (defined(__linux__) || defined(__FreeBSD__) || defined(__OpenBSD__)) && > \ > ++ defined(__aarch64__) > + #include <sys/auxv.h> // For getauxval() > + #endif > + > +@@ -179,7 +180,7 @@ LIBYUV_API SAFEBUFFERS int ArmCpuCaps(const char* cpui > + } > + > + #ifdef __aarch64__ > +-#ifdef __linux__ > ++#if defined(__linux__) || defined(__FreeBSD__) || defined(__OpenBSD__) > + // Define hwcap values ourselves: building with an old auxv header where > these > + // hwcap values are not defined should not prevent features from being > enabled. > + #define YUV_AARCH64_HWCAP_ASIMDDP (1 << 20) > +@@ -464,13 +465,20 @@ static SAFEBUFFERS int GetCpuFlags(void) { > + cpu_info |= kCpuHasLOONGARCH; > + #endif > + #if defined(__aarch64__) > +-#if defined(__linux__) > ++#if defined(__linux__) || defined(__FreeBSD__) || defined(__OpenBSD__) > ++#if defined(__FreeBSD__) || defined(__OpenBSD__) > ++ unsigned long hwcap = 0; > ++ unsigned long hwcap2 = 0; > ++ elf_aux_info(AT_HWCAP, &hwcap, sizeof(hwcap)); > ++ elf_aux_info(AT_HWCAP2, &hwcap2, sizeof(hwcap2)); > ++#else > + // getauxval is supported since Android SDK version 18, minimum at time of > + // writing is 21, so should be safe to always use this. If getauxval is > + // somehow disabled then getauxval returns 0, which will leave Neon > enabled > + // since Neon is mandatory on AArch64. > + unsigned long hwcap = getauxval(AT_HWCAP); > + unsigned long hwcap2 = getauxval(AT_HWCAP2); > ++#endif > + cpu_info = AArch64CpuCaps(hwcap, hwcap2); > + #else > + cpu_info = AArch64CpuCaps(); > Index: pkg/PLIST > =================================================================== > RCS file: /cvs/ports/graphics/libavif/pkg/PLIST,v > retrieving revision 1.6 > diff -u -p -u -p -r1.6 PLIST > --- pkg/PLIST 27 Aug 2024 07:54:17 -0000 1.6 > +++ pkg/PLIST 20 Mar 2025 01:58:40 -0000 > @@ -1,5 +1,6 @@ > @bin bin/avifdec > @bin bin/avifenc > +@bin bin/avifgainmaputil > include/avif/ > include/avif/avif.h > include/avif/avif_cxx.h >
