A faster more extensible radix sort

[LilithHafner]

Roadmap

• Comprehensive benchmarking
• Better heuristic for
  • radix size &
  • delegating to counting,
  • quick, and
  • merge sorts
• Ensure that any regressions in the extensible & adaptive final product compared to the original radix sort are warranted
• Update Base.Sort.defalg (if target is Base)
• Sort by and with (benchmark to see if these are worth including)
• Test policy on non-homogenous types

ref: https://discourse.julialang.org/t/request-for-comments-a-roadmap-to-faster-sort-for-primitive-types-using-radixsort/70269?

[codecov-commenter]

Codecov Report

Merging #53 (b7e6987) into master (d0a9007) will decrease coverage by 8.25%.
The diff coverage is 68.93%.

@@            Coverage Diff             @@
##           master      #53      +/-   ##
==========================================
- Coverage   97.84%   89.58%   -8.26%     
==========================================
  Files           1        4       +3     
  Lines         325      461     +136     
==========================================
+ Hits          318      413      +95     
- Misses          7       48      +41     

Impacted Files	Coverage Δ
src/SortingAlgorithms.jl	97.87% <ø> (+0.02%)	⬆️
src/radix.jl	52.94% <52.94%> (ø)
src/serialize.jl	70.90% <70.90%> (ø)
src/radix_wrapper.jl	96.15% <96.15%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d0a9007...b7e6987. Read the comment docs.

[LilithHafner]

High power testing via SortMark is incompatible with Julia 1.0. No intention to fix this rn.

[LilithHafner]

Closed in favor of JuliaLang/julia#44230