fuzzer: fix ivf parsing
fuzzer: use only external API and do not depend on config.h
fix -Wsign-compare warnings
meson: error on nasm < 2.13
API: always unref input buffer after consuming it completely
tile: check for arithmetic underflow in tile setup
parsing: error out when trying to decode a frame without tile data
Lossless segments: allow CfL in 4x4 chroma blocks
Move delta_lf inside delta_q check
Correctly use dav1d_ prefix for global symbols
x86: Enable ITX AVX2 asm on 64-bit Windows
x86: Fix out-of-bounds write in ITX asm
tests: add libfuzzer test target
decode_frame: unref everything on decoding errors
frame header: check for arithmetic underflow in tile data parsing
frame header: fix tile size parsing for non-uniform tiles
get_bits: avoid infinite loops in get_vlc() at EOF
x86-64: Inverse transforms AVX2
checkasm: Add inverse transform unit tests
build: Fix stack alignment
Remove dav1d_init()
Improve public API documentation in data.h
Improve public API documentation in dav1d.h
Remove compare between int and uint in add macro
attributes: fix clzll on 32-bit MSVC
decode: Fix unchecked dav1d_alloc_aligned in decode_frame
decode: Fix unchecked malloc in decode_frame
ref_mvs: Check malloc in av1_init_ref_mv_common and propagate error
lib: Fix unchecked malloc
ref_mvs: Check for alloation fail before calling memset in av1_alloc_ref_mv_common
Build: Do not warn about unused-parameter
Build: Do not warn about maybe-uninitialized
CI: added macOS build job.
lf_apply: explicitly mark a constant as unsigned
Change n_fc to unsigned
meson: Move pkg-config generation in src/
CI: add build-debian-aarch64 and build-debian-aarch64-clang-5
CI: Add a static build on debian
Build: Fix static library building
Fix version.h include
Remove compare between int and uint
Remove uninitialized fields in dav1d_cli_parse.c
Remove trailing spaces in y4m2.c
Remove unused parameters in get_tx_ctx
Fix ss_ver assignment in thread_picture_wait()
Fix for progress at edges of image
Fix occasional hang in tile row threading
build: use -Werror=vla to prevent variable length arrays
Windows: Fix conflicting pthread symbol names
Update README.md
aarch64: Always use the PIC version of movrel for iOS
checkasm: Define the default call_new if not using a checked_call wrapper
api: add a frame type field to Dav1dPictureParameters
Remove VLAs from Loop Restoration
thread: switch mutex wrappers to SRW locks
thread: Silence warning about unused attr parameter in pthread_create()
threads: Inline some Windows pthreads compat wrappers in the header
Build: Do not use soversion for windows dll
winver: Use already defined constants for FILETYPE and FILEOS
winver: Use dots in the VERSION_NUMBER_STR define
winver: Add missing required VERSIONINFO resource strings
Fix include order of config.h
Build: Link dav1d with static libdav1d on Windows
Add Macro to make it clear the variable is intended to not be initialized
build: add support for arm/aarch64 asm and integrate checkasm
add simplistic stdatomic.h compatible with MSVC
Remove DC row buffer for CfL prediction
Simplify conditions for CfL prediction
Introduce single-plane variant of cfl_pred
checkasm: Fix comparison of different int signs
checkasm: Fix missing-field-initializers warnings
Build: Add build dir to NASM include path
Set build warnings to 2 (-Wextra)
Build: Add missing license header
Build: Add option to disable building tests
Build: Add option to disable building tools
Build: Re-structure and cleanup meson.build files
Build: Move ext folder to src
Check for valid input in dav1d_data_unref
Workaround two more ubsan warnings.
x86: Enable MC AVX2 asm on 64-bit Windows
Fixed MemorySanitizer: use-of-uninitialized-value
Remove unneeded 'once' guards
Wrappable data structure
Build: Fix library version
Build: Fix thread dependency
checkasm: Add MC unit test
Add 'checkasm' asm testing/benchmarking framework
build: use local getopt for anything system where it's missing
CI: Do not run ninja with the verbose flag
Fix left shift of negative value -1
win32: Generate a resource file & link it with the library
meson: Extract version components
meson: Don't hardcode the version multiple times
gitlab-ci: Bump the image version