Rakefile

# Copyright 2011-2014 Brent Rowland.
# Use of this source code is governed the Apache License, Version 2.0, as described in the LICENSE file.

require 'iconv'

codepages = %w(
  ISO-8859-1 ISO-8859-2 ISO-8859-3 ISO-8859-4 ISO-8859-5 ISO-8859-6 ISO-8859-7 ISO-8859-8 ISO-8859-9 
  ISO-8859-10 ISO-8859-11 ISO-8859-13 ISO-8859-14 ISO-8859-15 ISO-8859-16 
  CP1252 CP1250 CP1251 CP1253 CP1254 CP1256 CP1257 CP1258 CP874)

codepages.each do |cp|
  desc "Generate codepage table for #{cp}"

  task cp do
    generate_codepage_table(cp)
  end

  desc "Generate codepage map for #{cp}"
  task "#{cp}_map" do
    generate_codepage_map(cp)
  end
end

task :maps => codepages.map { |cp| "#{cp}_map" }

# task :default => codepages + [:cpselect]

def generate_codepage_table(cp)
  puts "Generating codepage table for #{cp}"
  ranges = codepage_table(cp)
  open(cp.downcase+".go", "w") do |f|
    f.puts "package codepage", ""
    f.puts "var #{cp.gsub('-', '_')} = Codepage{"
    ranges.each do |range|
      f.puts "\t{0x%04X, 0x%04X, %d, %d}," % [range.first, range.first + range.count - 1, range.count, range.delta]
    end
    f.puts "}"
  end
end

def codepage_table(cp)
  codepoints = Iconv.open("UCS-2BE//IGNORE", cp) do |ic|
    (0..255).map { |c| ic.iconv(c.chr) }.map { |s| s.unpack('n') }.map { |a| a.first }
  end
  ranges_from_ary(codepoints, cp).sort! { |a, b| a.first <=> b.first }
end

def codepage_map(cp)
  Iconv.open("UCS-2BE//IGNORE", cp) do |ic|
    (0..255).map { |c| ic.iconv(c.chr) }.map { |s| s.unpack('n') }.map { |a| a.first }
  end
end

def generate_codepage_map(cp)
  puts "Generating codepage map for #{cp}"
  map = codepage_map(cp)
  open(cp.downcase+"_map.go", "w") do |f|
    f.puts "// Copyright 2014 Brent Rowland."
    f.puts "// Use of this source code is governed the Apache License, Version 2.0, as described in the LICENSE file."
    f.puts
    f.puts "package codepage", ""
    f.puts "var #{cp.gsub('-', '_')}_Map = []rune{"
    map.each_with_index do |r, i|
      f.puts "\t0x%04X, // #{i}" % [r, i]
    end
    f.puts "}"
  end
end

CPRange = Struct.new(:first, :count, :delta, :codepage)

class CPRanges < Array
  attr_accessor :name

  def char_for_codepoint(cp)
    low, high = 0, self.size - 1
    while low <= high
      i = (low + high) / 2
      r = self[i]
      if cp < r.first
        high = i - 1
        next
      end
      if cp >= r.first + r.count
        low = i + 1
        next
      end
      return cp + r.delta
    end
    return nil
  end
end

def ranges_from_ary(codepoints, name)
  ranges = CPRanges.new
  ranges.name = name
  range = nil
  last_codepoint = 0

  codepoints.each_with_index do |codepoint, i|
    next if codepoint.nil?
    if range.nil? or (codepoint != last_codepoint + 1)
      range = CPRange.new(codepoint, 1, i - codepoint)
      ranges << range
    else
      range.count += 1
    end
    last_codepoint = codepoint
  end

  ranges
end

class Codepages < Array
  def codepage_for_codepoint(cp)
    self.size.times do |i|
      if ch = self[i].char_for_codepoint(cp)
        # puts "#{cp}: #{ch}, #{self[i].name}"
        return self[i].name 
      end
    end
    nil
  end
end

desc "Generate codepage selection table"
task :cpselect do
  tables = Codepages.new.replace(codepages.map { |cp| codepage_table(cp) })
  codepoint_codepages = (0..0x2122).map { |cp| tables.codepage_for_codepoint(cp) }
  # codepoint_codepages.each_with_index { |codepage, codepoint| puts "#{codepoint}: #{codepage.inspect}" }
  ranges = []
  range = nil
  last_codepage = 0
  codepoint_codepages.each_with_index do |codepage, codepoint|
    next if codepage.nil?
    if range.nil? or (codepage != last_codepage)
      range = CPRange.new(codepoint, 1, nil, codepage)
      ranges << range
    else
      range.count += 1
    end
    last_codepage = codepage
  end
  open("cpselect.go", "w") do |f|
    f.puts "package codepage", ""

    f.puts "const ("
    codepages.each_with_index do |cp, i|
      f.puts "\tidx_%s%s"% [cp.gsub('-', '_'), i == 0 ? " = iota" : ""]
    end
    f.puts ")"
    f.puts
    f.puts "var CodepointCodepages = CodepageRanges{"
    ranges.each do |range|
      f.puts "\t{0x%04X, 0x%04X, %d, %s}," % [range.first, range.first + range.count - 1, range.count, "idx_" + range.codepage.gsub('-', '_')]
    end
    f.puts "}", ""

    f.puts "var Codepages = []Codepage{"
    codepages.each { |cp| f.puts "\t%s," % cp.gsub('-', '_') }
    f.puts "}"
  end
end