Fix Gitlab CI
cmake instructions
ci: add ci-fairy linter to make sure commits are GPG signed
ci: add arm64 CI
meson: fix build on arm64
docs: replace fgrep with grep -F
oops, avoid using a reserved identifier
Avoid "ISO C forbids an empty translation unit"
Make CELT FFT twiddle complex type aligned
opus.m4: fix -Wstrict-prototypes
Fix typo in MacroDebug.h comment.
Change pitch scaling behavior wrt nFramesPerPacket
Smooth out the LBRR rate estimate
More FEC tuning: lowering the LBRR bitrate a bit
Re-tuning the use of LTP scaling
Ensuring we can see where crashes occur
More ubsan fixes for the debug macros themselves
Using saturating round to fix some wrap-arounds
Relaxing checks for MULT16_32_QX()
Silence MSVC C4244 warning
Silence Clang 13+ null-pointer-subtraction warning
Only build platform RTCD sources when enabled.
Don't compile x86 cpu detection without RTCD.
Make silk/x86 header indentation consistent.
Fix C90-related warnings
update doc on custom mode
cmake - move warning C4244 to level 4
cmake - fix rtcd detection on x86 non windows
cmake - fix lrintf, lrint detection
Fix uninitialized field on custom mode malloc fail
meson: Fix reporting of cpu family if intrinsics not supported
Update x86 CPU detection configure check.
Fix warnings when compiling with FUZZING enabled
cmake - Add OPUS_BUILD to test targets
doc: Use consistent alternative notation
Fix opus.h for doxygen when referencing alternative values
Remove unused variable in tests
Replace assert with test_failed function in test
Fix warning with --disable-rfc8251
Fix quoting and whitespace errors in build test
Silence GCC 11+ -Wmaybe-uninitialized warnings
Avoids incrementing uninitialized values
Avoid undefined behaviour within the debug macros
Fix fixed-point overflow in pitch downsampling
Fix some 16-bit overflows (using 32-bit macros)
Avoid left shifts of negative values in debug macros
Work around UBSan unaligned access errors.
Add asan/ubsan support in random tests
Adds OPUS_SET_INBAND_FEC(2) option
Estimate the inner product accuracy to fix check-asm
Adds fuzzing to CPU detection
Work around a valgrind false-positive in CPUID.
Check the return value of __get_cpuid().
Fixes valgrind failure caused by silk_find_pred_coefs_*()
Silence some warnings for fixed-point debug builds
Fixes wrap-around in silk_inner_prod16_sse4_1()
Fix 8101b33 to decode ignored redundancy
Improve background noise estimation for CELT DTX
Correct redundancy handling with lost/DTX frames
print rate used for testvectors
Fix lrint/lrintf detection
Cleanup testing directories to save space
Update and re-enable SILK SSE4.1 optimisations
Prevent int32 overflow when applying HARM FIR filter in NSQ.c by using a saturating sum. This matches behavior in NSQ_del_dec.c.
Initialize non-zero test arrays.
Disable dangerous SSE 4.1 intrinsic optimizations
Fix buffer overflow in xcorr_kernel_sse4_1
Check channels/stream counts and mapping when creating the multistream
meson: fix get-version script for git worktrees
Revert relaxing comparison to 0 for fixed point only
Relax comparison to 0 to avoid a floating point divide-by-zero error.
cmake - add support to run ctest on android #2347
celt_lpc: avoid overflows when computing lpcs in fixed point
Sending refresh DTX packets every 400 ms independently of the encoded frame size.
Fix float-approx negative left shift UB
ci: fix pipeline run for merge requests
docs: fix simple typo, neareast -> nearest
Meson: Fix doc build when opus is a subproject
cmake - add option for fixed point debug
cmake - add option for check asm
cmake - add option for fuzzing
cmake - add option for hardening
cmake - add option for assertions
CMakeLists.txt: specify working directory in add_test() directives
CMakeLists.txt: specify path to target file in add_test() directives
cmake - change logging of api version to debug
Add support for Meson build system
Fix MSVC warning about trunction from double to float
repacketizer_demo: check for read errors to fix compiler warnings
Replace WIN32 with _WIN32 everywhere
celt: Fix broken SSE pre-processor check due to typo
silk: Fix incorrect ifdef in debug.c
opus_decoder_fuzzer: limit the number of decodes to avoid timeout
cmake - MINGW check for ssp lib and link if security features is enabled