Up to [cvs.NetBSD.org] / pkgsrc / audio / flac / patches
Request diff between arbitrary revisions
Keyword substitution: kv
Default branch: MAIN
audio/flac: Update to 1.5.0 Changelog: 1.5.0: Changes As there have been additions to the libFLAC interfaces, the libFLAC version number is incremented to 14. The libFLAC++ version number is incremented to 11. General Multithreaded encoding is now possible in libFLAC and through the flac command line tool The GFDL license file is updated to version 1.3 The markdown tool documentation is now also converted to HTML, for bundling with systems that do not read manpages (e.g. Windows) Decoding of chained Ogg FLAC files is now possible (philippe44, Martijn van Beurden) Various fixes (Sam James, Miroslav Lichvar, Cristian Rodríguez, manxorist, kgroeneveld, Lee Carré, Jevin Sweval, braheezy, Wolfgang Stöggl) Is is now possible in libFLAC, libFLAC++ and metaflac to write to a new file when changing metadata, instead of needing to overwrite an existing file flac Testing mode (flac -t) now parses all metadata blocks and warns the user when ID3v1 metadata is detected A warning is displayed when frame numbers do not increase correctly throughout a file The explain option (-H or --explain) is now removed, use the manpage or html tool documentation instead Built-in help and tool documentation are improved (H2Swine) When re-encoding a FLAC file from an existing FLAC file, a check is added that the MD5 sums of both files are the same libFLAC and libFLAC++ The library interfaces have been extended. See the porting guide (part of the API documentation) An error is sent when a frame is missing The algorithm of the 'loose mid side' option has changed. Instead of checking every few frames which option is best and keeping that for the next few frames, a fast heuristic is now used. This was necessary to enable multithreading Most level 0 metadata interface functions now also work with Ogg FLAC files When encoding Ogg FLAC files, the callback now returns a number of samples instead of always 0 (Jesper Larsson, ziplantil) When changing metadata, libFLAC now detects when an input file is a symlink, and will refuse to write data to it when an in-place rewrite of the metadata cannot happen When encoding using seektable templates, unused seekpoints (with a sample number higher than the total number of samples) are converted to placeholders Build system Fix building on Android with API version < 24 (Steve Lhomme) The microbench utility has been removed Enable building with emscripten (werner mendizabal) Minimum CMake version required (when building with CMake) is now formally 3.12 Testing/validation Improve fuzzing of allocation failures Various other fuzzing improvements Documentation The FLAC format is now specified in RFC 9639 The foreign metadata storage format used by the flac command line tool is now properly documented
flac: updated to 1.3.3 FLAC 1.3.3: General: - Fix CPU detection - Switch from unsigned types to uint32_t - CppCheck fixes - Improve SIMD decoding of 24 bit files - POWER* amnd POWER9 improvements - More tests. flac: - When converting to WAV, use WAVEFORMATEXTENSIBLE when bits per second is not 8 or 16 - Fix --output-prefix with input-files in sub-directories build system: - Cmake support - Visual Studio updates - Fix for MSVC when UNICODE is enabled - Fix for OpenBSD/i386
FLAC 1.3.2: General: * Fix undefined behaviour using GCC/Clang UBSAN (erikd). * General hardening via fuzz testing with AFL (erikd and others). * General code improvements (lvqcl, erikd and others). * Add FLAC in MP4 specification docs (Ralph Giles). * MSVS build cleanups (lvqcl). * Fix some cppcheck warnings (erikd). * Assume all currently used OSes support SSE2. flac: * Fix potential infinite loop on flac-to-flac conversion (erikd). * Add WAVEFORMATEXTENSIBLE to WAV (as needed) when decoding (lvqcl). * Only write vorbis-comments if they are non-empty. * Error out if decoding RAW with bits != (8|16|24). metaflac: * Add --scan-replay-gain option. build system: * Fixes for MSVC and Makefile.lite build systems. libraries: * CPU detection cleanup and fixes (Julian Calaby, erikd and lvqcl). * Fix two stream decoder bugs (Max Kellermann). * Fix a NULL dereference bug (on a malformed file). * Changed the LPC order guess for a slight compression improvement, particularly for classical music (Martijn van Beurden). * Improved encoding speed on older Intel CPUs. * Fixed a seeking bug when decoding certain files (Miroslav Lichvar). * Put an upper bound (32768) on the number of seek points. * Fix potential memory leaks. * Support 64bit brword/bwword allowing FLAC__BYTES_PER_WORD to be set to 8 (disabled by default). * Fix an out-of-bounds heap read. * Win32: Only use large buffers when writing to disk.
Don't remove -g, especially when it breaks e.g. -gdwarf-2. Add patch comment.