# Disable the automatically set build flags to the environment variables such # as CC, CFLAGS, and etc. # See . %undefine _auto_set_build_flags Name: simde Version: v0.8.4.rc3 Release: 1.20260909145407199299.pr1449.101.g761be0ef%{?dist} Summary: Dummy License: MIT URL: https://github.com/simd-everywhere/simde # This value is magically replaced with the archive tar.gz file created by # `git archive` in the process of the CI. # Set the existing file `README.md` in case of creating SRPM from this RPM spec # file on local for testing purpose. Source0: simde-v0.8.4.rc3.tar.gz # A path to the CI script. Source1: ci.sh # List up the needed RPM package names to test here. # clang, clang++ BuildRequires: clang BuildRequires: compiler-rt BuildRequires: gcc # g++ BuildRequires: gcc-c++ # ninja BuildRequires: ninja-build # pip and setuptools are needed to install and use meson PyPI package. BuildRequires: python3-devel BuildRequires: python3-pip BuildRequires: python3-setuptools # The GNU time (/bin/time) is used for testing. Use this GNU time rather than # the embedded time function in bash. # https://www.gnu.org/software/time/ BuildRequires: time # RPM package description. %description # The bash commands to prepare are executed. %prep # Print the replaced value of the Source0. echo "Source0: %{SOURCE0}" # Create a dummy file for a valid RPM spec file. touch dummy.txt # The bash commands to test are executed. %check # Check the directory structure in the tar.gz file. tar tzvf "%{SOURCE0}" > source_files.txt head -2 source_files.txt # Check if the submodule munit is included. grep munit source_files.txt # Extract the source archive file. tar xzf "%{SOURCE0}" pushd simde-* # Set the flags used to build the actual RPM package. # These can be used in the CI testing script. %global toolchain clang export CI_CLANG_RPM_CFLAGS="%{build_cflags}" export CI_CLANG_RPM_CXXFLAGS="%{build_cxxflags}" export CI_CLANG_RPM_LDFLAGS="%{build_ldflags}" %global toolchain gcc export CI_GCC_RPM_CFLAGS="%{build_cflags}" export CI_GCC_RPM_CXXFLAGS="%{build_cxxflags}" export CI_GCC_RPM_LDFLAGS="%{build_ldflags}" echo "Running the CI script %{SOURCE1}." /bin/time -f '=> [%E]' "%{SOURCE1}" popd # Files included in the built RPM package. %files %doc dummy.txt # Changelog entries for the RPM package. %changelog * Wed Sep 09 2026 Packit - v0.8.4.rc3-1.20260909145407199299.pr1449.101.g761be0ef - build(deps): bump hendrikmuhs/ccache-action from 1.2.23 to 1.2.24 (dependabot[bot]) - msvc arm64 neon: work around 2 miscompilation bugs (Michael R. Crusoe) - msvc: no support for 128-bit types (Michael R. Crusoe) - x86 avx512: fix for MSVC /arch:AVX10.x for BF16 (Michael R. Crusoe) - msvc: still no AVX512FP16, even with /arch:AVX10.1 or .2 (Michael R. Crusoe) - gh-actions: add native arm64 msvc (Michael R. Crusoe) - x86 avx10: turn on most SIMDE_ARCH_X86_AVX512* flags when AVX10.x is enabled (Michael R. Crusoe) - appveyor: msvc 2026, build for AVX10.[12] (Michael R. Crusoe) - gh-actions clang-qemu & macos: upgrade and test newer (Michael R. Crusoe) - gh-actions macos 14* is deprecated. (Michael R. Crusoe) - gh-actions: test clang 23 (Michael R. Crusoe) - [skip ci] gh-actions PR cache cleanup: tweak to enable correct secrets from forks (Michael R. Crusoe) - gh-actions: pr-cleanup needs a full runner for permissions reasons (Michael R. Crusoe) - gh-actions: test clang 22 (Michael R. Crusoe) - revert ubuntu upgrades that failed (Michael R. Crusoe) - gh-actions: switch to arm64 based qemu builds (Michael R. Crusoe) - gh-actions clang-qemu s390x fix (Michael R. Crusoe) - gh-actions ci: drop clang-19 on Intel due flakiness from https://github.com/actions/runner-images/issues/14296 (Michael R. Crusoe) - gh-actions: simplify how the correct libstdc++-*-dev-*-cross packages are selected (Michael R. Crusoe) - gh-actions: upgrade to Ubuntu 26.04 where possible (Michael R. Crusoe) - gcc arm neon bug fixed in gcc-17 (Michael R. Crusoe) - arm neon ld2 gcc: expand use of SIMDE_DIAGNOSTIC_DISABLE_UNINITIALIZED_ (Michael R. Crusoe) - SDE flag checks, implement workaround for pclmul, sse3, and -march=x86-64 (Michael R. Crusoe) - gh-actions: run cache cleanup on faster ubuntu-slim (Michael R. Crusoe) - x86 aes native aliases: add missing #undefs (Michael R. Crusoe) - x86 bmi: enable the 64-bit native intrinsics only for 64-bit systems. (Michael R. Crusoe) - azure: revert ubuntu-26.04 upgrade (Michael R. Crusoe) - azure: ignore branches targeting other CI systems (Michael R. Crusoe) - packit: ignore branches specific to other CI providers (Michael R. Crusoe) - circleci: new syntax, add retries for flaky arm64 Ubuntu mirror (Michael R. Crusoe) - gcc-16 -O3 silence uninitialized warning (Michael R. Crusoe) - gh-actions gcc-16: ubuntu 26.04 with newer qemu (10.2.1+) (Michael R. Crusoe) - gh-actions actionlint: upgrade to recognize ubuntu-26.04 (Michael R. Crusoe) - gh-actions: upgrade qemu for loong64 (Michael R. Crusoe) - gh-actions clang-qemu: test versions 21 & 22 for loong64 with -ffast-math (Michael R. Crusoe) - test: Avoid -ffinite-math-only on floating point comparisons (Ivan A. Melnikov) - gh-actions: test loongarch64 with -Ofast (Michael R. Crusoe) - CI: drop testing on ubuntu 22.04 (Michael R. Crusoe) - ci: add cross-compilation for Loongarch64 GCC 14 with only LSX enabled (Elysia) - x86/avx: Fix LSX implementation issues in comparison and load functions (Elysia) - Delete .github/workflows/ossar-analysis.yml (Michael R. Crusoe) - gh-actions: new PPA for ccache (Michael R. Crusoe) - gh-actions: safer ccache cache key construction (Michael R. Crusoe) - check for SIMDE_SHUFFLE_VECTOR_ when using it (Michael R. Crusoe) - build(deps): bump EndBug/add-and-commit from 11.0.0 to 11.1.1 (#1435) (dependabot[bot]) - build(deps): bump github/codeql-action/upload-sarif (#1436) (dependabot[bot]) - x86/gfni: performance improvements (fo40225) - gh-actions gcc-16: fix install (Michael R. Crusoe) - gh-actions: upgrade actions/checkout (Michael R. Crusoe) - x86/avx, avx2: use SIMDE_LOONGARCH_LSX_NATIVE guard for Loongarch LSX intrinsics (elysia) - x86/sse: fix conditional compilation for Loongarch LSX (Yinan Qin) - build(deps): bump EndBug/add-and-commit from 10.0.0 to 11.0.0 (dependabot[bot]) - build(deps): bump actions/setup-python from 6.2.0 to 7.0.0 (dependabot[bot]) - build(deps): bump actions/checkout from 6.0.2 to 7.0.1 (dependabot[bot]) - build(deps): bump zizmorcore/zizmor-action from 0.6.1 to 0.6.2 (dependabot[bot]) - gh-actions: drop codeql, no longer free of charge (Michael R. Crusoe) - gh-actions: zizmor fixes (Michael R. Crusoe) - build(deps): bump actions/setup-dotnet from 5.2.0 to 6.0.0 (dependabot[bot]) - build(deps): bump hendrikmuhs/ccache-action from 1.2.20 to 1.2.23 (dependabot[bot]) - gh-actions clang-21 x64 without march=native (Michael R. Crusoe) - gh-actions: macos 26.{4.1,5} with clang-21 (Michael R. Crusoe) - gh-actions: x86 sapphire rapids, include bf16 feature. (Michael R. Crusoe) - tests: add bfloat support; fix buffer overflow error (Michael R. Crusoe) - build(deps): bump zizmorcore/zizmor-action from 0.5.3 to 0.6.1 (dependabot[bot]) - x86/clmul: allow VEX-encoded _mm256_clmulepi64_epi128 without AVX512VL (fo4025) - simd128: fix AltiVec version guards for POWER8-only intrinsics (fo4025) - arm: fix AltiVec version guards for POWER8-only intrinsics (neon/paddl) (fo4025) - ci: add GCC 14 power7 cross job to gcc-qemu matrix (fo4025) - x86: fix AltiVec version guards for POWER8-only intrinsics (sse2, sse4.1) (fo4025) - math: use fpclassify for isinf on PowerPC to avoid QEMU xststdcsp bug (fo4025) - x86/avx512/cvt: add simde_mm512_cvtepu8_epi32 (fo4025) - gh-actions: deploy, adjust permissions (Michael R. Crusoe) - build(deps): bump ad-m/github-push-action from 1.1.0 to 1.3.0 (dependabot[bot]) - x86/avx512/compress: add maskz_compress_epi8 and mask_compressstoreu_epi8 (fo4025) - migrate Cirrus CI to CircleCI (Michael R. Crusoe) - gh-actions: move formatting check to stand-alone file for local usage (Michael R. Crusoe) - bug submitted and fixed upstream. (Michael R. Crusoe) - x86/avx512/dpbf16: avoid GCC's broken 512-bit masked intrinsics (fo4025) - tighten github actions with zizmor (Michael R. Crusoe) - x86/cvt: update conversion values to use INT32_MIN for edge cases (Felix Sternberg) - x86/avx2: fix conversion in cvt.h to use direct access (Felix Sternberg) - x86/avx2: add implementation and test for simde_mm256_maskz_loadu_epi8 (Felix Sternberg) - x86/avx512: fix macro definition for simde_mm512_mask_extracti32x8_epi32 to use correct feature guard (Felix Sternberg) - x86/avx512: add implementations and tests for simde_mm256_extracti64x2_epi64, simde_mm256_mask_extracti64x2_epi64, and simde_mm256_maskz_extracti64x2_epi64 (Felix Sternberg) - x86/avx512: fix simde_mm512_mask_extracti64x2_epi64 and simde_mm512_maskz_extracti64x2_epi64 to use simde_mm_mask_mov_epi64 and simde_mm_maskz_mov_epi64 (Felix Sternberg) - x86/avx2: fix cvt.h to use SIMDE_X86_AVX2_NATIVE guard (Felix Sternberg) - x86/avx512: add implementations and tests for simde_mm512_extracti64x2_epi64, simde_mm512_mask_extracti64x2_epi64, and simde_mm512_maskz_extracti64x2_epi64 (Felix Sternberg) - x86/avx2: fix cvt.h to use HEDLEY_REINTERPRET_CAST and HEDLEY_STATIC_CAST for type conversions (Felix Sternberg) - x86/avx512: add implementations and tests for simde_mm_sllv_epi16 and simde_mm256_sllv_epi16 (Felix Sternberg) - x86/avx512: add simde_mm_cvtepi16_epi8 implementation and tests (Felix Sternberg) - x86/avx512: add cvt intrinsics in cvt.h with matching tests in cvt.c (Felix Sternberg) - add implementations and tests for simde_mm_cmpgt_epi8_mask, simde_mm256_cmpgt_epi8_mask (Felix Sternberg) - build(deps): bump ad-m/github-push-action from 1.0.0 to 1.1.0 (dependabot[bot]) - math: extend non-glibc guard to Clang and add exp10 guard (Jeongkeun Kim) - ci: add Alpine (musl + GCC + SIMDE_NO_NATIVE) to Cirrus CI (jg.kim0724) - math: restrict __builtin_roundeven to glibc >= 2.25 on GCC 10+ (jg.kim0724) - build(deps): bump EndBug/add-and-commit from 9 to 10 (#1397) (dependabot[bot]) - x86 avx512: add more reduction functions for 32-bit and 64-bit integers, floating-point, and bitwise operations (#1395) (Adrian Riedl) - semaphore ci: removed, they no longer support Ubuntu 20.04 (Michael R. Crusoe) - x86 avx: fix cvtt bounds (Max Slater) - various fixes inspired by gcc -fanalyzer (Michael R. Crusoe) * Fri Jun 02 2023 Dummy Dummy - 1-1 - Dummy.