pixman: Update to version 0.46.0

pixman: Update to version 0.46.0

- Update from version 0.44.0 to 0.46.0
- Update of rootfile
- Changelog
    0.46.0
	This release notably adds fast paths for RISC-V using the "V" vector
	 extension, contributed by developers at Samsung.
	      RISC-V: Only enable RVV on linux if hwcap headers are available
	      Test case for compositing with a negative stride
	      Fix arm64 advanced prefetcher
	      region: add parametric primitive type to generalize implementation detail
	      region: make print specifier parametric
	      region: add fractional implementation based on 64bit floating point numbe
	      region: add rectf convenience functions
	      region: add pixman_region32_copy_from_region64f utility function
	      region: add image clip and composite functions for fractional regions
	      test/region: add fractional region tests
	      Add a16b16g16r16 format
	      test/stress-test: add a16b16g16r16
	      ci: Improve coverage and artifact handling
	      ci: Use newer version of QEMU for Bookworm
	      test: Increase timeout for alpha-loop test
	      ci: Enable cross PPC tests
	      ci: Clean up after disabling failing targets
	      ci: Add option to change the default runner tag
	      ci: Move Docker build to a CI template
	      ci: Use native ARM runners for Linux ARM targets
	      ci: Improve Windows images
	      ci: Use regular Debian as base Windows for ARM
	      ci: Fix runner tag variable expansion
	      Fix some build warning
	      rvv: Pass through clang-format
	      docker: Update Wine to 10.5 and update LLVM-MinGW
	      docker: Update Meson to 1.7
	      docker: Move LLVM archive to a separate thin layer
	      docker: Decrease build time for gcovr
	      test: Add lowlevel-blt-bench result comparison script
	      rvv: Add float suffix to float implementation
	      rvv: Add integer operations
	      ci: Build Wine for windows-amd64 from source
	      ci: Use untagged runners if possible
	      ci: Increase number of retries for Windows targets
	      Post-release version bump to 0.44.3
	      vmx: Remove unnecessary variable
	      vmx: Remove unpack_565_to_8888() and associated constants
	      vmx: Remove unpack_128_2x128_16()
	      vmx: Remove unpack{hi,lo}_128_8x16 functions
	      vmx: Move and use unpack{hi,lo}_128_16x8 in pix_multiply()
	      vmx: Use create_mask_32_128() more places
	      vmx: Use appropriate types
	      vmx: Add and use create_mask_16_128() function
	      vmx: Use selector variables
	      vmx: Add and use `vzero` constant
	      vmx: Simplify over() function
	      vmx: Make in_over() a real function
	      vmx: Simplify unpack{hi,lo}_128_16x8() function
	      vmx: Move create_mask_32_128() function
	      vmx: Avoid two shifts in pix_multiply() function
	      vmx: Optimize pix_multiply()
	      vmx: Declare iterator variable in for loop
	      vmx: Disable clang-format around some data structure declarations
	      vmx: Run clang-format
	      ci: Enable LLVM testing on linux-arm-v7
	      ci: Allow failures in windows-amd64 jobs
	      Pre-release version bump to 0.46.0
	      arm: Move the .fpu neon directive around
	      .clang-format: Add
    0.44.2
	      meson: require Meson >= 1.3.0
	      ci: Rename mips64el to mips64le
	      ci: Enable ppc64le testing with clang
	      Post-release version bump to 0.44.1
	      vmx: Fix is_opaque, is_zero, is_transparent functions
	      release.sh: Some improvements
	      Pre-release version bump to 0.44.2
	      RISC-V: fix rvv auto-detection on `gcc-13`
	      release.sh: Add script
	      RISC-V: Force spec 1p0 in architecture definiton

Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>