fix hang in dav1d_close() with more than 31 tile threads
tools/dav1d: don't call write_header and write_trailer when it's not needed
Add a null muxer which will just drop the packet.
libfuzzer: unbreak oss-fuzz build
rename arch specific bitdepth template files
Build: Add suffix to templated BITDEPTH files
meson: add "-Werror=missing-prototypes" to the list of optional arguments
attributes: Don't align to more than what assembly needs/benefits from
fix unwanted integer promotion in tile data size parsing.
loop restoration: maximal stripe height is 128 + 8
Don't call mask_edges_chroma when !bw4 or !bh4
attributes: Fix leftover mismatched else/endif comments
arm/mc: Add 8 bit neon asm for avg, w_avg and mask
Make access into loopfilter masks safe across sb64 tile threads
Make access into noskip_mask safe across sb64 tile threads
Make symbols shared with libaom static.
Convert PATENTS file from CP1252 to UTF-8.
Prevent right-shift by 32
Revert "Add tile constraint to loopfilter mask fix-up code"
fix debug build after 16ec1057342
obu: don't abort OBU_REDUNDANT_FRAME_HDR if no frame header has been parsed
obu: abort parsing OBU_FRAME if show_existing_frame=1
CI: add -Werror to all warning free jobs and add clang-7 job
remove unused function av1_clamp_mv()
Don't fix chroma loopfilter masks for grayscale content
Simplify code to fix-up the tile boundary loopfilter masks
Add tile constraint to loopfilter mask fix-up code
Exit frame data decoding for OBU_FRAME if show_existing_frame=1
Fix compiler warning in checkasm/looprestoration.c
Initialize restoration.type to RESTORATION_NONE for grayscale content
Use max value for all trailing palette entries
Require refs and curpic to share the same chroma subsampling
libfuzzer: fix a -Wformat warning
CI: add a -Werror build without checkasm
fix sign compare warnings in dav1d_loopfilter_sbrow
arm64: mc: Make the jump tables local symbols
remove unused unused code from src/ref_mvs.c
obu: support parsing Redundant Frame Header OBUs
fuzzer: add a multithreaded fuzzer using 2 frame and tile threads
fuzzer: add a standalone fuzzing engine 'none'
fuzzer: improve coverage of lib.c
fuzzer: prevent unwanted integer promotion in r32le
obu: fix parsing of decoder model info bits
getbits: do not call get_bits(0) for coded value 0 in dav1d_get_vlc
Fix negative shift exponent in get_poc_diff
Avoid calling get_uniform(max=1)
Order skipmode references incrementally
Clamp bottom end of vertical tile length
Rename unset_hp_bit() to fix_mv_precision() and add integer-mv support.
Add --demuxer CLI parameter.
Use aligned image edges for intrabc
Tiny refactoring to make intra_edge more readable
Fix palette values when we reach max plane values
obu: fix parsing force_integer_mv on frame headers
arm64: Don't use uxth for extending a register
Fill refpoc even if use_ref_frame_mvs=0.
Update txpart/inter_txtp tables for intra frames also.
Fix looprestoration intermediate upper limit
arm64/mc: add 8-bit neon asm for avg, w_avg and mask
looprestoration/x86: use x86 dsp init only if HAVE_ASM is 1
x86: fix redefinition of _WIN32_WINNT in looprestoration
x86: Add filter intra prediction AVX2 asm
Reorder the dav1d_filter_intra_taps array
x86: Add pal_pred AVX2 asm
obu/tile_group: simplify tile group order to check
ivf: fix the conversion specifier in printed error message strings
Fix wiener overflow (#79).
meson: Disable asserts in release builds
Fix wiener AVX2 SIMD and related unit test.
Add support for external picture buffer allocation
x86: Add chroma-from-luma intra prediction AVX2 asm
checkasm: Add unit tests for chroma-from-luma
Move chroma-from-luma alpha to last parameter
meson: Define CONFIG_xBPC to 0 for the bitdepths that aren't enabled
attributes: Don't try to use _BitScanReverse64 on MSVC/ARM
checkasm: Fix building with some bitdepths disabled
checkasm: Include windows.h for the MSVC ARM64 readtime macro
data: remove unnecessary internal header
fix skip_mode_params parsing if enable_order_hint == 0
parse obu_size/leb128 as unsigned 32-bit integer
dav1d/ivf: return an error when fread fails
dav1d_decode: avoid broken and unecessary int cast in assert
check segmap of primary reference frame before referencing it
intrabc: clip displacement vector to current tile
decode: allow returning an error from decode_b and pass it on
Add infrastructure for LR SIMD and unit tests.
Refactor left edge copying to reduce data copies by 50%.
Small clean up in loop filter code
Defer DC_PRED for zero-alpha CFL_PRED
Remove cfl_uvtx and cfl_uv_t_dim
Compute DC within chroma-from-luma prediction
Remove dual-plane chroma-from-luma prediction
x86: Add smooth intra prediction AVX2 asm
free tile_thread task_idx_to_sby_and_tile_idx
implement non power of 2 tile rows and columns spec compliant
tools/dav1d: unref input data after decoding loop
decode: verify inter frame size restrictions
checkasm: Don't use readtime for checkasm seed