Simple ctags generator for ruby code using melbourne.

1-usage.md

$ ruby tags.rb
    5	def	Rubinius.ruby18?
    6	class	SyntaxError
    7	def	SyntaxError.from
   13	class	Rubinius::AST::Node
   14	def	Rubinius::AST::Node#visit
   26	def	Rubinius::AST::Node#full_name
   31	class	Visitor
   32	def	Visitor#initialize
   46	def	Visitor#emit_tag
   54	def	Visitor#s_class
   62	def	Visitor#s_class_end
   69	def	Visitor#class
   81	def	Visitor#class_end
   87	def	Visitor#define
   98	def	Visitor#define_end
  102	def	Visitor#define_singleton
  106	def	Visitor#define_singleton_end
  111	def	Visitor#constant_assignment

$ ruby tags.rb /usr/lib/ruby/1.8/timeout.rb
   30	module	Timeout
   35	class	Timeout::Error
   37	class	Timeout::ExitException
   40	const	Timeout::THIS_FILE
   41	const	Timeout::CALLER_OFFSET
   52	def	Timeout#timeout
  100	def	Object#timeout
  108	const	TimeoutError

$ ruby tags.rb /usr/lib/ruby/1.8/timeout.rb --ctags
!_TAG_FILE_FORMAT	2	/extended format/
!_TAG_FILE_SORTED	0	/0=unsorted, 1=sorted, 2=foldcase/
Timeout	/usr/lib/ruby/1.8/timeout.rb	/^module Timeout$/;"	m
Error	/usr/lib/ruby/1.8/timeout.rb	/^  class Error < Interrupt$/;"	c	class:Timeout	inherits:Interrupt
ExitException	/usr/lib/ruby/1.8/timeout.rb	/^  class ExitException < ::Exception # :nodoc:$/;"	c	class:Timeout	inherits:Exception
THIS_FILE	/usr/lib/ruby/1.8/timeout.rb	/^  THIS_FILE = \/\\A#{Regexp.quote(__FILE__)}:\/o$/;"	C
CALLER_OFFSET	/usr/lib/ruby/1.8/timeout.rb	/^  CALLER_OFFSET = ((c = caller[0]) && THIS_FILE =~ c) ? 1 : 0$/;"	C
timeout	/usr/lib/ruby/1.8/timeout.rb	/^  def timeout(sec, klass = nil)$/;"	f	class:Timeout
timeout	/usr/lib/ruby/1.8/timeout.rb	/^def timeout(n, e = nil, &block) # :nodoc:$/;"	f
TimeoutError	/usr/lib/ruby/1.8/timeout.rb	/^TimeoutError = Timeout::Error # :nodoc:$/;"	C

$ ruby tags.rb /usr/lib/ruby/1.8/timeout.rb --json
{"kind":"module","language":"Ruby","line":30,"full_name":"Timeout","name":"Timeout","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"module Timeout"}
{"kind":"class","language":"Ruby","line":35,"full_name":"Timeout::Error","name":"Error","class":"Timeout","inherits":"Interrupt","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"  class Error < Interrupt"}
{"kind":"class","language":"Ruby","line":37,"full_name":"Timeout::ExitException","name":"ExitException","class":"Timeout","inherits":"Exception","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"  class ExitException < ::Exception # :nodoc:"}
{"kind":"constant","language":"Ruby","line":40,"name":"THIS_FILE","full_name":"Timeout::THIS_FILE","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"  THIS_FILE = /\\A#{Regexp.quote(__FILE__)}:/o"}
{"kind":"constant","language":"Ruby","line":41,"name":"CALLER_OFFSET","full_name":"Timeout::CALLER_OFFSET","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"  CALLER_OFFSET = ((c = caller[0]) && THIS_FILE =~ c) ? 1 : 0"}
{"kind":"method","language":"Ruby","line":52,"name":"timeout","class":"Timeout","full_name":"Timeout#timeout","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"  def timeout(sec, klass = nil)"}
{"kind":"method","language":"Ruby","line":100,"name":"timeout","full_name":"Object#timeout","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"def timeout(n, e = nil, &block) # :nodoc:"}
{"kind":"constant","language":"Ruby","line":108,"name":"TimeoutError","full_name":"TimeoutError","path":"/usr/lib/ruby/1.8/timeout.rb","pattern":"TimeoutError = Timeout::Error # :nodoc:"}

tags.rb

require 'melbourne'

if RUBY_ENGINE == 'ruby'
  def Rubinius.ruby18?() true end
  class SyntaxError
    def self.from(*args)
      new(args.first)
    end
  end
end

class Rubinius::AST::Node
  def visit(visitor, parent=nil)
    if visitor.respond_to?(self.node_name)
      visitor.__send__ self.node_name, self, parent
    end

    children { |c| c.visit visitor, self }

    if visitor.respond_to?("#{self.node_name}_end")
      visitor.__send__ "#{self.node_name}_end", self, parent
    end
  end

  def full_name
    respond_to?(:parent) && parent ? "#{parent.full_name}.#{name}" : name.to_s
  end
end

class Visitor
  def initialize(ast = nil)
    @namespace = []
    @definition = []

    @tags = []
    ast.visit(self) if ast
  end

  attr_reader :tags
  undef send

  ###
  # helpers

  def emit_tag(kind, node, opts={})
    opts.delete_if{ |k,v| v.nil? }
    @tags << {:kind => kind.to_s, :language => 'Ruby', :line => node.line}.merge(opts)
  end

  ###
  # visitor callbacks

  def s_class(node, parent)
    if node.receiver.node_name != 'self'
      @namespace << node.receiver.full_name
    end

    @prev_is_singleton = @is_singleton
    @is_singleton = true
  end
  def s_class_end(node, parent)
    if node.receiver.node_name != 'self'
      @namespace.pop
    end
    @is_singleton = @prev_is_singleton
  end

  def class(node, parent)
    @namespace << node.name.full_name
    superclass = node.respond_to?(:superclass) && node.superclass

    parts = @namespace.join('.').split('.')
    name  = parts.pop
    emit_tag node.node_name, node,
      :full_name => @namespace.join('.').gsub('.','::'),
      :name => name,
      :class => parts.any? && parts.join('.') || nil,
      :inherits => superclass.respond_to?(:name) && superclass.full_name || nil
  end
  def class_end(node, parent)
    @namespace.pop
  end
  alias module class
  alias module_end class_end

  def define(node, parent)
    @definition << node
    return if @definition.size > 1 # skip nested defines

    emit_tag @is_singleton ? 'singleton method' : 'method', node,
      :name => node.name.to_s,
      :class => @namespace.any?? @namespace.join('.') : nil,
      :full_name =>
        (@namespace.any? ? @namespace.join('.').gsub('.','::') : 'Object') +
        (@is_singleton ? ".#{node.name}" : "##{node.name}")
  end
  def define_end(node, parent)
    @definition.pop
  end

  def define_singleton(node, parent)
    s_class(node, parent)
    define(node.body, parent)
  end
  def define_singleton_end(node, parent)
    define_end(node.body, parent)
    s_class_end(node, parent)
  end

  def constant_assignment(node, parent)
    emit_tag :constant, node,
      :name => node.constant.name.to_s,
      :full_name => [@namespace + [node.constant.full_name]].join('::')
  end
end

if ARGV.delete('--json')
  require 'yajl'
  json = true
end
if ARGV.delete('--ctags')
  ctags = true
  puts "!_TAG_FILE_FORMAT\t2\t/extended format/"
  puts "!_TAG_FILE_SORTED\t0\t/0=unsorted, 1=sorted, 2=foldcase/"
end

ARGV.unshift __FILE__ if ARGV.empty?
ARGV.each do |file|
  begin
    if ast = Rubinius::Melbourne.parse_file(file)
      lines = File.readlines(file)
      v = Visitor.new(ast)
      v.tags.each do |tag|
        tag[:path] = file
        tag[:pattern] = lines[tag[:line]-1].chomp

        if json
          puts Yajl.dump(tag)

        elsif ctags
          kwargs = ''
          kwargs << "\tclass:#{tag[:class]}" if tag[:class]
          kwargs << "\tinherits:#{tag[:inherits]}" if tag[:inherits]

          kind = case tag[:kind]
          when 'method' then 'f'
          when 'singleton method' then 'F'
          when 'constant' then 'C'
          else tag[:kind].slice(0,1)
          end

          code = tag[:pattern].gsub('\\','\\\\\\\\').gsub('/','\\/')
          puts "%s\t%s\t/^%s$/;\"\t%c%s" % [tag[:name], file, code, kind, kwargs]

        else
          kind = case tag[:kind]
          when /method$/ then 'def'
          when /^const/  then 'const'
          else tag[:kind]
          end

          puts "#{tag[:line].to_s.rjust(5)}\t#{kind}\t#{tag[:full_name]}"
        end
      end
    end
  rescue SyntaxError
    STDERR.puts [$!, file].inspect
  end
end