class StringPrepTablesGenerator

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 19
def initialize(rfc_filename:  STRINGPREP_RFC_FILE,
               json_filename: STRINGPREP_JSON_FILE)
  @rfc_filename  = rfc_filename
  @json_filename = json_filename
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 42
def arrays;  @arrays  ||= ranges.transform_values{|t| t.flat_map(&:to_a) } end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 45
def asgn_regexps; @asgn_regexps || asgn_regexps! end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 28
def clean_deps; Rake::FileList.new           STRINGPREP_JSON_FILE end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 30
def generate_json_data_file
  require "json"
  rfc_filename
    .then(&File.method(:read))
    .then(&method(:parse_rfc_text))
    .then(&JSON.method(:pretty_generate))
    .then {|data| File.write json_filename, data }
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 26
def json_deps;  Rake::FileList.new __FILE__, STRINGPREP_RFC_FILE  end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 47
def merged_tables_regex(*table_names, negate: false)
  table_names
    .flat_map(&arrays.method(:fetch))
    .then {|array| to_regexp(array, negate: negate) }
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 41
def ranges;  @ranges  ||= tables.transform_values(&method(:to_ranges)) end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 27
def rb_deps;    Rake::FileList.new __FILE__, STRINGPREP_JSON_FILE end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 53
def regexp_for(*names, negate: false)
  asgn_regexps[[*names, negate]] ||= merged_tables_regex(*names, negate: negate)
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 44
def regexps; @regexps ||= arrays.transform_values(&method(:to_regexp)) end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 152
  def saslprep_rb
    <<~RUBY
      # frozen_string_literal: true

      #--
      # This file is generated from RFC3454, by rake.  Don't edit directly.
      #++

      module Net::IMAP::SASL

        module SASLprep

          # RFC4013 §2.1 Mapping - mapped to space
          # * non-ASCII space characters (\\StringPrep\\[\\"C.1.2\\"]) that can be
          #   mapped to SPACE (U+0020), and
          #
          # Equal to \\StringPrep\\[\\"C.1.2\\"].
          # Redefined here to avoid loading the StringPrep module.
          MAP_TO_SPACE = #{regex_str "C.1.2"}

          # RFC4013 §2.1 Mapping - mapped to nothing
          #   the "commonly mapped to nothing" characters (\\StringPrep\\[\\"B.1\\"])
          #   that can be mapped to nothing.
          #
          # Equal to \\StringPrep\\[\\"B.1\\"].
          # Redefined here to avoid loading the StringPrep module.
          MAP_TO_NOTHING = #{regex_str "B.1"}

          # RFC4013 §2.3 Prohibited Output::
          # * Non-ASCII space characters — \\StringPrep\\[\\"C.1.2\\"]
          # * ASCII control characters — \\StringPrep\\[\\"C.2.1\\"]
          # * Non-ASCII control characters — \\StringPrep\\[\\"C.2.2\\"]
          # * Private Use characters — \\StringPrep\\[\\"C.3\\"]
          # * Non-character code points — \\StringPrep\\[\\"C.4\\"]
          # * Surrogate code points — \\StringPrep\\[\\"C.5\\"]
          # * Inappropriate for plain text characters — \\StringPrep\\[\\"C.6\\"]
          # * Inappropriate for canonical representation characters — \\StringPrep\\[\\"C.7\\"]
          # * Change display properties or deprecated characters — \\StringPrep\\[\\"C.8\\"]
          # * Tagging characters — \\StringPrep\\[\\"C.9\\"]
          TABLES_PROHIBITED = #{SASL_TABLES_PROHIBITED.inspect}.freeze

          # Adds unassigned (by Unicode 3.2) codepoints to TABLES_PROHIBITED.
          #
          # RFC4013 §2.5 Unassigned Code Points::
          #   This profile specifies the \\StringPrep\\[\\"A.1\\"] table as its list of
          #   unassigned code points.
          TABLES_PROHIBITED_STORED = ["A.1", *TABLES_PROHIBITED].freeze

          # Matches codepoints prohibited by RFC4013 §2.3.
          #
          # See TABLES_PROHIBITED.
          #
          # Equal to +Regexp.union+ of the TABLES_PROHIBITED tables.  Redefined
          # here to avoid loading the StringPrep module unless necessary.
          PROHIBITED_OUTPUT = #{regex_str(*SASL_TABLES_PROHIBITED)}

          # RFC4013 §2.5 Unassigned Code Points::
          #   This profile specifies the \\StringPrep\\[\\"A.1\\"] table as its list of
          #   unassigned code points.
          UNASSIGNED = #{regex_str "A.1"}

          # Matches codepoints prohibited by RFC4013 §2.3 and §2.5.
          #
          # See TABLES_PROHIBITED_STORED.
          PROHIBITED_OUTPUT_STORED = Regexp.union(
            UNASSIGNED, PROHIBITED_OUTPUT
          ).freeze

          # Bidirectional Characters [StringPrep, §6]
          BIDI_FAILURE = #{bidi_failure_regexp.inspect}.freeze

          # Matches strings prohibited by RFC4013 §2.3 and §2.4.
          #
          # This checks prohibited output and bidirectional characters.
          PROHIBITED = Regexp.union(
            PROHIBITED_OUTPUT, BIDI_FAILURE,
          )

          # Matches strings prohibited by RFC4013 §2.3, §2.4, and §2.5.
          #
          # This checks prohibited output, bidirectional characters, and
          # unassigned codepoints.
          PROHIBITED_STORED = Regexp.union(
            PROHIBITED_OUTPUT_STORED, BIDI_FAILURE,
          )

        end
      end
    RUBY
  end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 43
def sets;    @sets    ||= arrays.transform_values(&:to_set) end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 57
  def stringprep_rb
    <<~RUBY
      # frozen_string_literal: true

      #--
      # This file is generated from RFC3454, by rake.  Don't edit directly.
      #++

      module Net::IMAP::SASL

        module StringPrep

          #{asgn_table "A.1"}

          #{asgn_table "B.1"}

          #{asgn_table "B.2"}

          #{asgn_table "B.3"}

          #{asgn_table "C.1.1"}

          #{asgn_table "C.1.2"}

          #{asgn_table "C.2.1"}

          #{asgn_table "C.2.2"}

          #{asgn_table "C.3"}

          #{asgn_table "C.4"}

          #{asgn_table "C.5"}

          #{asgn_table "C.6"}

          #{asgn_table "C.7"}

          #{asgn_table "C.8"}

          #{asgn_table "C.9"}

          #{asgn_table "D.1"}

          # Used to check req3 of bidirectional checks
          #{asgn_table "D.1", negate: true}

          #{asgn_table "D.2"}

          BIDI_DESC_REQ2 = "A string with RandALCat characters must not contain LCat characters."

          # Bidirectional Characters [StringPrep, §6], Requirement 2::
          #   If a string contains any RandALCat character, the string MUST NOT
          #   contain any LCat character.
          BIDI_FAILS_REQ2 = #{bidi_fails_req2.inspect}.freeze

          BIDI_DESC_REQ3 = "A string with RandALCat characters must start and end with RandALCat characters."

          # Bidirectional Characters [StringPrep, §6], Requirement 3::
          #   If a string contains any RandALCat character, a RandALCat
          #   character MUST be the first character of the string, and a
          #   RandALCat character MUST be the last character of the string.
          BIDI_FAILS_REQ3 = #{bidi_fails_req3.inspect}.freeze

          # Bidirectional Characters [StringPrep, §6]
          BIDI_FAILURE = #{bidi_failure_regexp.inspect}.freeze

          # Names of each codepoint table in the RFC-3454 appendices
          TABLE_TITLES = {
            #{table_titles_rb}
          }.freeze

          # Regexps matching each codepoint table in the RFC-3454 appendices
          TABLE_REGEXPS = {
            #{table_regexps_rb}
          }.freeze

        end
      end
    RUBY
  end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 145
def table_regexps_rb(indent = 3)
  asgn_regexps # => { ["A.1", false] => regexp, ... }
    .reject {|(_, n), _| n }
    .map {|(t, _), _| "%p => %s," % [t, regexp_const_name(t)] }
    .join("\n#{"  "*indent}")
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 139
def table_titles_rb(indent = 3)
  titles
    .map{|t| "%p => %p," % t }
    .join("\n#{"  "*indent}")
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 39
def tables;  @tables  ||= load_tables_and_titles_from_json!.first end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 40
def titles;  @titles  ||= load_tables_and_titles_from_json!.last end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 365
def asgn_regex(name, regexp, negate: false)
  asgn_regexps[[name, negate]] = regexp
  "# %s\n%s%s = %p.freeze" % [
    regexp_const_desc(name, negate: negate), " " * 4,
    regexp_const_name(name, negate: negate),
    regexp,
  ]
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 309
def asgn_regexps!
  @asgn_regexps = {}
  # preset the regexp for each table
  asgn_regex "A.1", /\p{^AGE=3.2}/
  # If ruby supported all unicode properties (i.e. line break = word joiner):
  #   /[\u{00ad 034f 1806}\p{join_c}\p{VS}\p{lb=WJ}&&\p{age=3.2}]/
  asgn_table "B.1"
  asgn_table "B.2"
  asgn_table "B.3"
  asgn_regex "C.1.1", / /
  asgn_regex "C.1.2", /[\u200b\p{Zs}&&[^ ]]/
  asgn_regex "C.2.1", /[\x00-\x1f\x7f]/
  # C.2.2 is a union:
  #   Cc + Cf (as defined by Unicode 3.2) + Zl + Zp + 0xfffc
  #   - any codepoints covered by C.2.1 or C.8 or C.9
  #
  # But modern Unicode properties are significantly different, so it's better
  # to just load the table definition.
  asgn_table "C.2.2"
  asgn_regex "C.3", /\p{private use}/
  asgn_regex "C.4", /\p{noncharacter code point}/
  asgn_regex "C.5", /\p{surrogate}/
  asgn_regex "C.6", /[\p{in specials}&&\p{AGE=3.2}&&\p{^NChar}]/
  asgn_regex "C.7", /[\p{in ideographic description characters}&&\p{AGE=3.2}]/
  # C.8 is a union of \p{Bidi Control} and Unicode 3.2 properties.  But those properties
  # have changed for modern Unicode, and thus for modern ruby's regexp
  # character properties.  It's better to just load the table definition.
  asgn_table "C.8"
  asgn_regex "C.9", /[\p{in Tags}&&\p{AGE=3.2}]/
  # Unfortunately, ruby doesn't (currently) support /[\p{Bidi
  # Class=R}\p{bc=AL}]/.  On the other hand, StringPrep (based on Unicode 3.2)
  # might not be a good match for the modern (14.0) property value anyway.
  asgn_table "D.1"
  asgn_table "D.1", negate: true # used by BIDI_FAILS_REQ3
  asgn_table "D.2"
  @asgn_regexps
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 351
def asgn_table(name, negate: false)
  asgn_regex(name, regexp_for(name, negate: negate), negate: negate)
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 376
def bidi_L        ; regexp_for "D.2" end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 374
def bidi_R_AL     ; regexp_for "D.1" end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 406
  def bidi_consts
    <<~RUBY
      #############
          # Bidirectional checks.
          #

    RUBY
  end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 378
def bidi_fails_req2
  / # RandALCat followed by LCat
    (?<r_and_al_cat>#{bidi_R_AL.source})
      .*?
      (?<l_cat>#{bidi_L.source})
    | # RandALCat preceded by LCat
      \g<l_cat> .*? \g<r_and_al_cat>
  /mux
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 388
def bidi_fails_req3
  / # contains RandALCat but doesn't start with RandALCat
    \A(?<not_r_nor_al>#{bidi_not_R_AL})
      .*?
      (?<r_and_al_cat>#{bidi_R_AL})
    | # contains RandALCat but doesn't end with RandALCat
      \g<r_and_al_cat> .*? \g<not_r_nor_al>\z
  /mux
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 399
def bidi_failure_regexp
  req3_with_backref = bidi_fails_req3.source
    .gsub(%r{\(\?\<r_and_al_cat\>\(.*?\)\)}, "\g<r_and_al_cat>")
    .then{|re|"(?mx-i:#{re})"}
  /#{bidi_fails_req2} | #{req3_with_backref}/mux
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 375
def bidi_not_R_AL ; regexp_for "D.1", negate: true end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 272
def load_tables_and_titles_from_json!
  require "json"
  @tables = json_filename
    .then(&File.method(:read))
    .then(&JSON.method(:parse))
  @titles = @tables.delete "titles"
  [@tables, @titles]
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 245
def parse_rfc_text(rfc3454_text)
  titles = {}
  tables, = rfc3454_text
    .lines
    .each_with_object([]) {|line, acc|
      current, table = acc.last
      case line
      when /^([A-D]\.[1-9](?:\.[1-9])?) (.*)/
        titles[$1] = $2
      when /^ {3}-{5} Start Table (\S*)/
        acc << [$1, []]
      when /^ {3}-{5} End Table /
        acc << [nil, nil]
      when /^ {3}([0-9A-F]+); ([ 0-9A-F]*)(?:;[^;]*)$/  # mapping tables
        table << [$1, $2.split(/ +/)] if current
      when /^ {3}([-0-9A-F]+)(?:;[^;]*)?$/              # regular tables
        table << $1 if current
      when /^ {3}(.*)/
        raise "expected to match %p" % $1 if current
      end
    }
    .to_h.compact
    .transform_values {|t| t.first.size == 2 ? t.to_h : t }
  tables["titles"] = titles
  tables
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 347
def regex_str(*names, negate: false)
  "%p.freeze" % regexp_for(*names, negate: negate)
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 355
def regexp_const_desc(name, negate: false)
  if negate then "Matches the negation of the %s table" % [name]
  else %q{%s \\StringPrep\\[\\"%s\\"]} % [titles.fetch(name), name]
  end
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 361
def regexp_const_name(table_name, negate: false)
  "IN_%s%s" % [table_name.tr(".", "_"), negate ? "_NEGATED" : ""]
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 281
def to_ranges(table)
  (table.is_a?(Hash) ? table.keys : table)
    .map{|range| range.split(?-).map{|cp| Integer cp, 16} }
    .map{|s,e| s..(e || s)}
end

# File net-imap-0.3.4/rakelib/string_prep_tables_generator.rb, line 289
def to_regexp(codepoints, negate: false)
  codepoints
    .grep_v(SURROGATES_RANGE) # remove surrogate codepoints from C.5 and D.2
    .uniq
    .sort
    .chunk_while {|cp1,cp2| cp1 + 1 == cp2 }     # find contiguous chunks
    .map {|chunk| chunk.map{|cp| "%04x" % cp } } # convert to hex strings
    .partition {|chunk| chunk[1] }               # ranges vs singles
    .then {|ranges, singles|
      singles.flatten!
      [
        negate ? "^" : "",
        singles.flatten.any? ? "\\u{%s}" % singles.join(" ") : "",
        ranges.map {|r| "\\u{%s}-\\u{%s}" % [r.first, r.last] }.join,
        codepoints.any?(SURROGATES_RANGE) ? "\\p{Cs}" : "", # not necessary :)
      ].join
    }
    .then {|char_class| Regexp.new "[#{char_class}]" }
end