v5.3.0

What's Changed

• Faster in-place union of a few containers into many by @lemire in #908
• vectorize-array-to-uint32 (faster array_container_to_uint32_array on ARM) by @SalvatorePreviti in #903
• Negate sparse arrays into run containers; SIMD in-place array union by @lemire in #906
• perf: keep the AVX-512 union reverse off the carried chain by @Dr-Emann in #900
• perf: fold the AVX-512 compare-exchange blend into a masked min by @Dr-Emann in #901
• perf: use a perfect-shuffle bitonic merge for the SSE union and xor by @lemire in #909
• Add roaring_bitmap_word_range_to_bitset by @lemire in #907

Full Changelog: v5.2.4...v5.3.0