Refactor to a regex-based algorithm

[skryukov]

The primary goal of this PR is to enhance the speed of URI::IDNA::UTS46, bringing it up to par with the speed of the pure IDNA algorithm from Addressable. Achieving this could enable URI::IDNA::UTS46 to be used as a default IDNA-algorithm in Addressable without significant performance trade-offs.

To accomplish this, this PR introduces several changes to the internal implementation algorithm—it includes additional checks for ascii_only?, as suggested in the ruby/ruby#8966 for Ruby, and replaces loops over characters/code points with regular expressions, among other minor enhancements. Additionally. This PR also includes more unit-tests for validation logic, yay 🎉

To assess the performance improvements, I used a modified version of the addressable benchmark:

# /usr/bin/env ruby
# frozen_string_literal: true.

require "benchmark"
require "uri/idna"
require "addressable"
require "addressable/idna/pure"

simple = "google.com"
unicode = "fiᆵリ宠퐱卄.com"
punycode = "xn--fi-w1k207vk59a3qk9w9r.com"
N = 100_000

Benchmark.bmbm do |x|
  x.report("pure") { N.times { Addressable::IDNA::Pure.to_unicode(Addressable::IDNA::Pure.to_ascii(simple)) } }
  x.report("pure unicode") { N.times { Addressable::IDNA::Pure.to_unicode(Addressable::IDNA::Pure.to_ascii(unicode)) } }
  x.report("pure punycode") { N.times { Addressable::IDNA::Pure.to_unicode(Addressable::IDNA::Pure.to_ascii(punycode)) } }

  x.report("uri/idna") { N.times { URI::IDNA.to_unicode(URI::IDNA.to_ascii(simple)) } }
  x.report("uri/idna unicode") { N.times { URI::IDNA.to_unicode(URI::IDNA.to_ascii(unicode)) } }
  x.report("uri/idna punycode") { N.times { URI::IDNA.to_unicode(URI::IDNA.to_ascii(punycode)) } }
end

Results (MacBook Pro 13", M1, 2020, 16GB, ruby v3.2.2):

Rehearsal -----------------------------------------------------
pure                0.447617   0.003119   0.450736 (  0.453697)
pure unicode        4.123272   0.006671   4.129943 (  4.133575)
pure punycode       1.982393   0.002661   1.985054 (  1.987714)
uri/idna            0.773894   0.000712   0.774606 (  0.775704)
uri/idna unicode    3.867693   0.001920   3.869613 (  3.874109)
uri/idna punycode   5.087081   0.002383   5.089464 (  5.095687)
------------------------------------------- total: 16.299416sec

                        user     system      total        real
pure                0.443630   0.000954   0.444584 (  0.445314)
pure unicode        4.185360   0.007045   4.192405 (  4.198606)
pure punycode       2.000301   0.003139   2.003440 (  2.006636)
uri/idna            0.779944   0.000704   0.780648 (  0.781443)
uri/idna unicode    3.888171   0.002455   3.890626 (  3.902315)
uri/idna punycode   5.334571   0.026223   5.360794 (  5.613100)