File rav1d.changes of Package rav1d
-------------------------------------------------------------------
Sat Aug 16 08:07:26 UTC 2025 - aloisio@gmx.com
- Update to version v1.1.0:
* Bump version numbers to 1.1.0 (#1389)
* `#[warn(static_mut_refs)]`: fix with an `#[allow(static_mut_refs)]` since it's in already `unsafe` tool code
* `#[warn(unpredictable_function_pointer_comparisons)]`: fix with `fn_addr_eq`
* `rust-toolchain.toml`: update to `nightly-2025-05-01` from `nightly-2024-06-13`
* arm: looprestoration: Fix the single line loop in sgr_weighted2
* arm: looprestoration: Add spacing around operators
* arm: looprestoration: Get rid of unnecessary rotate_ab_N intermediate functions
* arm: looprestoration: Apply 'const' more consistently on parameters
* arm: Use /proc/cpuinfo on linux if getauxval is unavailable
* Port C code changes to Rust
* AArch64: Add Neon implementation of load_tmvs
* arm64: looprestoration: Fix a comment typo
* arm64: looprestoration: Fix use of the wrong register
* Port C code changes to Rust
* arm64: looprestoration: Remove an unnecessary duplicate parameter in dav1d_sgr_weighted2_Xbpc_neon
* arm64: looprestoration: Apply simplifications to align with C code
* arm: looprestoration: Split an overly long line
* Rearrange default cdf values to enable fewer copies
* cdf: Combine memcpy() calls in dav1d_cdf_thread_copy()
* Add TODO item regarding buffer layout
* x86: Eliminate hardcoded struct offsets in refmvs load_tmvs() asm
* Port C code changes to Rust
* x86: Use the decl and init macros for put_8tap and prep_8tap
* x86: Improve SSSE3 SGR asm
* Add building and calling of ARM64 SVE2 assembly functions
* Fix the macro parameter name for the CHECK_SIZE macro
* Port C code changes to Rust
* Ensure that the refmvs_refpair union is packed
* `build.rs`: update to match changes in `meson.build`
* Rust: Remove references to deleted 8bpc mc SSE2 asm
* x86: Make AVX2 SGR gatherless
* x86: Add 6-tap variants of high bit-depth mc SSSE3 functions
* x86: Add 6-tap variants of 8bpc mc SSSE3 functions
* x86: Add minor 8bpc mc SSE improvements
* x86: Remove 8bpc mc SSE2 asm
* x86: Remove unused macro in mc16_avx512.asm
* AArch64: Trim Armv8.0 Neon path of 6-tap and 8-tap MC functions
* AArch64: New method for calculating sgr table
* AArch64: Optimize lane load/store in MC functions
* AArch64: Optimize Armv8.0 Neon path of SBD H/HV 6-tap filters
* AArch64: Optimize Armv8.0 Neon path of HBD HV 6-tap filters
* AArch64: Optimize Armv8.0 Neon path of HBD horizontal 6-tap filters
* AArch64: Optimize Armv8.0 Neon path of HBD horizontal filters
* aarch64: Split the jump tables to a separate const section
* aarch64: Enable detection of SVE/SVE2 on Windows
* aarch64: Fix a label typo
* aarch64: Avoid looping through the BTI instructions
* aarch64: ipred: Use the right fill width loop in ipred_z3_fill_padding_neon
* AArch64: SVE MS armasm64 fix of HBD subpel filters
* aarch64: mc16: Optimize the BTI landing pads in put/prep_neon
* AArch64: Add HBD subpel filters using 128-bit SVE2
* AArch64: Add USMMLA impl. for SBD 6-tap H/HV filters
* AArch64: Fix typo in SBD 6-tap 2D/HV subpel filter
* AArch64: Move constants of DotProd subpel filters to .rodata
* aarch64: Explicitly use the ldur instruction where relevant in mc_dotprod.S
* Add Arm OpenBSD run-time CPU feature detection support
* Use an older ubuntu image in CI (#1383)
* refmvs: Consolidate r and rp_proj allocations
* `ObuMetaType`: fix enum name for ITU-T T.35 metadata
* Port C code changes to Rust
* refmvs: Remove dav1d_refmvs_init()
* refmvs: Simplify 2-pass logic
* arm32/msac: Trim C functions, saves 1024 bytes
* arm64/msac: Trim C functions, saves 1392 bytes
* Correct ISRG attribution
* Verify alignment for AlignedVec
* `build.rs`: Add `#[cfg(debug_assertions)]` to `-F`/`-f` flags because they imply `-g` as well.
* `tools/build.rs`: Only run when compiling for `*-windows-msvc`, and error if trying to cross compile to `*-windows-msvc`.
* `tools/build.rs`: Move `rav1d-lib` specific stuff to its own `build.rs` so `rav1d` can cross compile to/from windows.
* `build.rs`: Make `-Fdwarf`/`-fwin64` flag based on target os, not host os.
* Fix bug: BT2020_CL vs BT2020_NC
* `fn DisjointMutArcSlice::from_iter`: Use inline `const {}` block to remove warning about dummy type.
* `fn mem::transmute`: Replace with sound `Arc::from_raw(Arc::into_raw(...) as *const _)` as `Arc` is not `#[repr(transparent)]`.
* `fn AlignedVec::resize`: Use `.expect` of `let else panic!`.
* Rewrite AligvedVec::resize check to checked_mul
* Address stylistic review comments
* Validate AlignedVec::resize safety requirements
* Fix length print when encountering an unknown OBU
* Make default logger stderr
* `fn cdef::Fn::call`: Use `.wrapping_as_ptr` for `bottom` to avoid out-of-bounds, as the ptr is temporarily out of bounds when passed to asm.
* `trait Pixels`: Add `fn wrapping_*as_{mut_,}ptr` methods for non-bounds checked versions.
* Make dotprod and i8mm configurable features
* `enum Av1BlockIntraInter`: Remove `#[repr(C)]`, reducing size from 28 to 24, and `Av1Block` from 36 to 32.
-------------------------------------------------------------------
Fri Sep 13 11:59:12 UTC 2024 - Luigi Baldoni <aloisio@gmx.com>
- Initial package (v1.0.0)