class Capybara::Selector::RegexpDisassembler

def initialize(regexp)
  @regexp = regexp
  @regexp_source = regexp.source
end

def simplify_group(matches)
  if matches[:group].include?('|') # no support for alternation in groups
    '.'
  elsif matches[:one_or_more] # required but may repeat becomes text + wildcard
    matches[:group][1..-2] + '.'
  elsif matches[:optional] # optional group becomes wildcard
    '.'
  elsif matches[:min_rep]
    (matches[:group] * matches[:min_rep].to_i).tap { |r| r << '.' if matches[:max_rep] }
  else
    matches[:group][1..-2]
  end
end

def substrings
  @substrings ||= begin
    source = @regexp_source.dup
    source.gsub!(/\\[^pgk]/, '.') # replace escaped characters with wildcard
    source.gsub!(/\\[gk](?:<[^>]*>)?/, '.') # replace sub expressions and back references with wildcard
    source.gsub!(/\\p\{[[:alpha:]]+\}?/, '.') # replace character properties with wildcard
    source.gsub!(/\[\[:[a-z]+:\]\]/, '.') # replace posix classes with wildcard
    while source.gsub!(/\[(?:[^\[\]]+)\]/, '.'); end # replace character classes with wildcard
    source.gsub!(/\(\?<?[=!][^)]*\)/, '') # remove lookahead/lookbehind assertions
    source.gsub!(/\(\?(?:<[^>]+>|>|:)/, '(') # replace named, atomic, and non-matching groups with unnamed matching groups
    while source.gsub!(GROUP_REGEX) { |_m| simplify_group(Regexp.last_match) }; end
    source.gsub!(/.[*?]\??/, '.') # replace optional character with wildcard
    source.gsub!(/(.)\+\??/, '\1.') # replace one or more with character plus wildcard
    source.gsub!(/(?<char>.)#{COUNTED_REP_REGEX.source}/) do |_m| # repeat counted characters
      (Regexp.last_match[:char] * Regexp.last_match[:min_rep].to_i).tap { |str| str << '.' if Regexp.last_match[:max_rep] }
    end
    return [] if source.include?('|') # can't handle alternation here
    strs = source.match(/\A\^?(.*?)\$?\Z/).captures[0].split('.').reject(&:empty?).uniq
    strs = strs.map(&:upcase) if @regexp.casefold?
    strs
  end
end