relax: std, core, and x86_64 SIMD intrinsics for test & windowed GUI closure
- Add x86_64 runtime feature detection for cmpxchg16b, movbe, xsave, fxsr, adx, rdrand, and rdseed in std/arch.rs (fearless_simd, pulp).
- Add _MM_FROUND_NO_EXC, _CMP_* constants, and 37 SSE4.2/AVX2 intrinsics in core/arch/x86_64.rs (fearless_simd, tiny_skia, vello_cpu, epaint).
- Add all 12 integer RHS shift impls (Shl, Shr, ShlAssign, ShrAssign) in core/num/prim_traits.rs and gen_num.rs, plus Wrapping<T> ShlAssign/ShrAssign<usize> in wrapping.rs (zmij, skrifa).
- Add core::convert::FloatToInt and f32/f64::to_int_unchecked (num-traits), lossless NonZero<T> From conversions, and unsigned Div/Rem/DivAssign/RemAssign<NonZero<T>> (rav1e, tiny_skia).
- Replace fixed-size smallvec::Array impls with const-generic unsafe impl<T, const N: usize> Array for [T; N], and add SmallVec::new_const and ArrayVec::new_const (idna, kurbo, avif-serialize).
- Add Vec::push_mut, VecDeque::{pop_front_if, pop_back_if, push_back_mut, push_front_mut}, Layout::pad_to_align, ptr::slice_from_raw_parts(_mut), NonNull::slice_from_raw_parts, [T; N]::each_mut, tuple-array From conversions, [u8]::escape_ascii, iter::repeat_n, and From<TryReserveError> for io::Error.
7bb2047f05