Soks 0.0.2

data/contrib/diff/lcs/hunk.rb

@@ -0,0 +1,257 @@
+#! /usr/env/bin ruby
+#--
+# Copyright 2004 Austin Ziegler <diff-lcs@halostatue.ca>
+#   adapted from:
+#     Algorithm::Diff (Perl) by Ned Konz <perl@bike-nomad.com>
+#     Smalltalk by Mario I. Wolczko <mario@wolczko.com>
+#   implements McIlroy-Hunt diff algorithm
+#
+# This program is free software. It may be redistributed and/or modified under
+# the terms of the GPL version 2 (or later), the Perl Artistic licence, or the
+# Ruby licence.
+# 
+# $Id: hunk.rb,v 1.2 2004/12/15 22:07:19 tamc Exp $
+#++
+# Contains Diff::LCS::Hunk for bin/ldiff.
+
+require 'diff/lcs/block'
+
+  # A Hunk is a group of Blocks which overlap because of the context
+  # surrounding each block. (So if we're not using context, every hunk will
+  # contain one block.) Used in the diff program (bin/diff).
+class Diff::LCS::Hunk
+    # Create a hunk using references to both the old and new data, as well as
+    # the piece of data
+  def initialize(data_old, data_new, piece, context, file_length_difference)
+      # At first, a hunk will have just one Block in it
+    @blocks = [ Diff::LCS::Block.new(piece) ]
+    @data_old = data_old
+    @data_new = data_new
+
+    before = after = file_length_difference
+    after += @blocks[0].diff_size
+    @file_length_difference = after # The caller must get this manually
+
+      # Save the start & end of each array. If the array doesn't exist
+      # (e.g., we're only adding items in this block), then figure out the
+      # line number based on the line number of the other file and the
+      # current difference in file lengths.
+    if @blocks[0].remove.empty?
+      a1 = a2 = nil
+    else
+      a1 = @blocks[0].remove[0].position
+      a2 = @blocks[0].remove[-1].position
+    end
+
+    if @blocks[0].insert.empty?
+      b1 = b2 = nil
+    else
+      b1 = @blocks[0].insert[0].position
+      b2 = @blocks[0].insert[-1].position
+    end
+
+    @start_old = a1 || (b1 - before)
+    @start_new = b1 || (a1 + before)
+    @end_old   = a2 || (b2 - after)
+    @end_new   = b2 || (a2 + after)
+
+    self.flag_context = context
+  end
+
+  attr_reader :blocks
+  attr_reader :start_old, :start_new
+  attr_reader :end_old, :end_new
+  attr_reader :file_length_difference
+
+    # Change the "start" and "end" fields to note that context should be added
+    # to this hunk
+  attr_accessor :flag_context
+  def flag_context=(context) #:nodoc:
+    return if context.nil? or context.zero?
+
+    add_start = (context > @start_old) ? @start_old : context
+    @start_old -= add_start
+    @start_new -= add_start
+
+    if (@end_old + context) > @data_old.size
+      add_end = @data_old.size - @end_old
+    else
+      add_end = context
+    end
+    @end_old += add_end
+    @end_new += add_end
+  end
+
+  def unshift(hunk)
+    @start_old = hunk.start_old
+    @start_new = hunk.start_new
+    blocks.unshift(*hunk.blocks)
+  end
+
+    # Is there an overlap between hunk arg0 and old hunk arg1? Note: if end
+    # of old hunk is one less than beginning of second, they overlap
+  def overlaps?(hunk = nil)
+    return nil if hunk.nil?
+
+    a = (@start_old - hunk.end_old) <= 1
+    b = (@start_new - hunk.end_new) <= 1
+    return (a or b)
+  end
+
+  def diff(format)
+    case format
+    when :old
+      old_diff
+    when :unified
+      unified_diff
+    when :context
+      context_diff
+    when :ed
+      self
+    when :reverse_ed, :ed_finish
+      ed_diff(format)
+    else
+      raise "Unknown diff format #{format}."
+    end
+  end
+
+  def each_old(block)
+    @data_old[@start_old .. @end_old].each { |e| yield e }
+  end
+
+  private
+    # Note that an old diff can't have any context. Therefore, we know that
+    # there's only one block in the hunk.
+  def old_diff
+    warn "Expecting only one block in an old diff hunk!" if @blocks.size > 1
+    op_act = { "+" => 'a', "-" => 'd', "!" => "c" }
+
+    block = @blocks[0]
+
+      # Calculate item number range. Old diff range is just like a context
+      # diff range, except the ranges are on one line with the action between
+      # them.
+    s = "#{context_range(:old)}#{op_act[block.op]}#{context_range(:new)}\n"
+      # If removing anything, just print out all the remove lines in the hunk
+      # which is just all the remove lines in the block.
+    @data_old[@start_old .. @end_old].each { |e| s << "< #{e}\n" } unless block.remove.empty?
+    s << "---\n" if block.op == "!"
+    @data_new[@start_new .. @end_new].each { |e| s << "> #{e}\n" } unless block.insert.empty?
+    s
+  end
+
+  def unified_diff
+      # Calculate item number range.
+    s = "@@ -#{unified_range(:old)} +#{unified_range(:new)} @@\n"
+
+      # Outlist starts containing the hunk of the old file. Removing an item
+      # just means putting a '-' in front of it. Inserting an item requires
+      # getting it from the new file and splicing it in. We splice in
+      # +num_added+ items. Remove blocks use +num_added+ because splicing
+      # changed the length of outlist.
+      #
+      # We remove +num_removed+ items. Insert blocks use +num_removed+
+      # because their item numbers -- corresponding to positions in the NEW
+      # file -- don't take removed items into account.
+    lo, hi, num_added, num_removed = @start_old, @end_old, 0, 0
+
+    outlist = @data_old[lo .. hi].collect { |e| e.gsub(/^/, ' ') }
+
+    @blocks.each do |block|
+      block.remove.each do |item|
+        op = item.action.to_s # -
+        offset = item.position - lo + num_added
+        outlist[offset].gsub!(/^ /, op.to_s)
+        num_removed += 1
+      end
+      block.insert.each do |item|
+        op = item.action.to_s # +
+        offset = item.position - @start_new + num_removed
+        outlist[offset, 0] = "#{op}#{@data_new[item.position]}"
+        num_added += 1
+      end
+    end
+
+    s << outlist.join("\n")
+  end
+
+  def context_diff
+    s = "***************\n"
+    s << "*** #{context_range(:old)} ****\n"
+    r = context_range(:new)
+
+      # Print out file 1 part for each block in context diff format if there
+      # are any blocks that remove items
+    lo, hi = @start_old, @end_old
+    removes = @blocks.select { |e| not e.remove.empty? }
+    if removes
+      outlist = @data_old[lo .. hi].collect { |e| e.gsub(/^/, '  ') }
+      removes.each do |block|
+        block.remove.each do |item|
+          outlist[item.position - lo].gsub!(/^ /) { block.op } # - or !
+        end
+      end
+      s << outlist.join("\n")
+    end
+
+    s << "\n--- #{r} ----\n"
+    lo, hi = @start_new, @end_new
+    inserts = @blocks.select { |e| not e.insert.empty? }
+    if inserts
+      outlist = @data_new[lo .. hi].collect { |e| e.gsub(/^/, '  ') }
+      inserts.each do |block|
+        block.insert.each do |item|
+          outlist[item.position - lo].gsub!(/^ /) { block.op } # + or !
+        end
+      end
+      s << outlist.join("\n")
+    end
+    s
+  end
+
+  def ed_diff(format)
+    op_act = { "+" => 'a', "-" => 'd', "!" => "c" }
+    warn "Expecting only one block in an old diff hunk!" if @blocks.size > 1
+
+    if format == :reverse_ed
+      s = "#{op_act[@blocks[0].op]}#{context_range(:old)}\n"
+    else
+      s = "#{context_range(:old).gsub(/,/, ' ')}#{op_act[@blocks[0].op]}\n"
+    end
+
+    unless @blocks[0].insert.empty?
+      @data_new[@start_new .. @end_new].each { |e| s << "#{e}\n" }
+      s << ".\n"
+    end
+    s
+  end
+
+    # Generate a range of item numbers to print. Only print 1 number if the
+    # range has only one item in it. Otherwise, it's 'start,end'
+  def context_range(mode)
+    case mode
+    when :old
+      s, e = (@start_old + 1), (@end_old + 1)
+    when :new
+      s, e = (@start_new + 1), (@end_new + 1)
+    end
+
+    (s < e) ? "#{s},#{e}" : "#{e}"
+  end
+
+    # Generate a range of item numbers to print for unified diff. Print
+    # number where block starts, followed by number of lines in the block
+    # (don't print number of lines if it's 1)
+  def unified_range(mode)
+    case mode
+    when :old
+      s, e = (@start_old + 1), (@end_old + 1)
+    when :new
+      s, e = (@start_new + 1), (@end_new + 1)
+    end
+
+    length = e - s + 1
+    first = (length < 2) ? e : s # "strange, but correct"
+    (length == 1) ? "#{first}" : "#{first},#{length}"
+  end
+end

data/contrib/diff/lcs/ldiff.rb

@@ -0,0 +1,226 @@
+#!/usr/bin/env ruby
+
+require 'optparse'
+require 'ostruct'
+require 'diff/lcs/hunk'
+
+  # == ldiff Usage
+  #   ldiff [options] oldfile newfile
+  #
+  # -c::                            Displays a context diff with 3 lines of context.
+  # -C [LINES], --context [LINES]:: Displays a context diff with LINES lines of context. Default 3 lines.
+  # -u::                            Displays a unified diff with 3 lines of context.
+  # -U [LINES], --unified [LINES]:: Displays a unified diff with LINES lines of context. Default 3 lines.
+  # -e::                            Creates an 'ed' script to change oldfile to newfile.
+  # -f::                            Creates an 'ed' script to change oldfile to newfile in reverse order.
+  # -a, --text::                    Treats the files as text and compares them line-by-line, even if they do not seem to be text.
+  # --binary::                      Treats the files as binary.
+  # -q, --brief::                   Reports only whether or not the files differ, not the details.
+  # --help::                        Shows the command-line help.
+  # --version::                     Shows the version of Diff::LCS.
+  #
+  # By default, runs produces an "old-style" diff, with output like UNIX diff.
+  #
+  # == Copyright
+  # Copyright © 2004 Austin Ziegler
+  #
+  #   Part of Diff::LCS <http://rubyforge.org/projects/ruwiki/>
+  #   Austin Ziegler <diff-lcs@halostatue.ca>
+  #
+  # This program is free software. It may be redistributed and/or modified under
+  # the terms of the GPL version 2 (or later), the Perl Artistic licence, or the
+  # Ruby licence.
+module Diff::LCS::Ldiff
+  BANNER = <<-COPYRIGHT
+ldiff #{Diff::LCS::VERSION}
+  Copyright � 2004 Austin Ziegler
+
+  Part of Diff::LCS.
+  http://rubyforge.org/projects/ruwiki/
+
+  Austin Ziegler <diff-lcs@halostatue.ca>
+
+  This program is free software. It may be redistributed and/or modified under
+  the terms of the GPL version 2 (or later), the Perl Artistic licence, or the
+  Ruby licence.
+
+$Id: ldiff.rb,v 1.2 2004/12/15 22:07:19 tamc Exp $
+              COPYRIGHT
+  
+  class << self
+    attr_reader   :format, :lines       #:nodoc:
+    attr_reader   :file_old, :file_new  #:nodoc:
+    attr_reader   :data_old, :data_new  #:nodoc:
+
+    def run(args, input = $stdin, output = $stdout, error = $stderr) #:nodoc:
+      args.options do |o|
+        o.banner = "Usage: #{File.basename($0)} [options] oldfile newfile"
+        o.separator ""
+        o.on('-c',
+            'Displays a context diff with 3 lines of',
+            'context.') do |ctx|
+          @format = :context
+          @lines  = 3
+        end
+        o.on('-C', '--context [LINES]', Numeric,
+            'Displays a context diff with LINES lines',
+            'of context. Default 3 lines.') do |ctx|
+          @format = :context
+          @lines  = ctx || 3
+        end
+        o.on('-u',
+            'Displays a unified diff with 3 lines of',
+            'context.') do |ctx|
+          @format = :unified
+          @lines  = 3
+        end
+        o.on('-U', '--unified [LINES]', Numeric,
+            'Displays a unified diff with LINES lines',
+            'of context. Default 3 lines.') do |ctx|
+          @format = :unified
+          @lines  = ctx || 3
+        end
+        o.on('-e',
+            'Creates an \'ed\' script to change',
+            'oldfile to newfile.') do |ctx|
+          @format = :ed
+        end
+        o.on('-f',
+            'Creates an \'ed\' script to change',
+            'oldfile to newfile in reverse order.') do |ctx|
+          @format = :reverse_ed
+        end
+        o.on('-a', '--text',
+             'Treat the files as text and compare them',
+             'line-by-line, even if they do not seem',
+             'to be text.') do |txt|
+          @binary = false
+        end
+        o.on('--binary',
+             'Treats the files as binary.') do |bin|
+          @binary = true
+        end
+        o.on('-q', '--brief',
+             'Report only whether or not the files',
+             'differ, not the details.') do |ctx|
+          @format = :report
+        end
+        o.on_tail('--help', 'Shows this text.') do
+          error << o
+          return 0
+        end
+        o.on_tail('--version', 'Shows the version of Diff::LCS.') do
+          error << BANNER
+          return 0
+        end
+        o.on_tail ""
+        o.on_tail 'By default, runs produces an "old-style" diff, with output like UNIX diff.'
+        o.parse!
+      end
+
+      unless args.size == 2
+        error << args.options
+        return 127
+      end
+
+        # Defaults are for old-style diff
+      @format ||= :old
+      @lines  ||= 0
+
+      file_old, file_new = *ARGV
+
+      case @format
+      when :context
+        char_old = '*' * 3
+        char_new = '-' * 3
+      when :unified
+        char_old = '-' * 3
+        char_new = '+' * 3
+      end
+
+        # After we've read up to a certain point in each file, the number of
+        # items we've read from each file will differ by FLD (could be 0).
+      file_length_difference = 0
+
+      if @binary.nil? or @binary
+        data_old = IO::read(file_old)
+        data_new = IO::read(file_new)
+
+          # Test binary status
+        if @binary.nil?
+          old_txt = data_old[0...4096].grep(/\0/).empty?
+          new_txt = data_new[0...4096].grep(/\0/).empty?
+          @binary = (not old_txt) or (not new_txt)
+          old_txt = new_txt = nil
+        end
+
+        unless @binary
+          data_old = data_old.split(/\n/).map! { |e| e.chomp }
+          data_new = data_new.split(/\n/).map! { |e| e.chomp }
+        end
+      else
+        data_old = IO::readlines(file_old).map! { |e| e.chomp }
+        data_new = IO::readlines(file_new).map! { |e| e.chomp }
+      end
+
+        # diff yields lots of pieces, each of which is basically a Block object
+      if @binary
+        diffs = (data_old == data_new)
+      else
+        diffs = Diff::LCS.diff(data_old, data_new)
+        diffs = nil if diffs.empty?
+      end
+
+      return 0 unless diffs
+
+      if (@format == :report) and diffs
+        output << "Files #{file_old} and #{file_new} differ\n"
+        return 1
+      end
+
+      if (@format == :unified) or (@format == :context)
+        ft = File.stat(file_old).mtime.localtime.strftime('%Y-%m-%d %H:%M:%S %z')
+        puts "#{char_old} #{file_old}\t#{ft}"
+        ft = File.stat(file_new).mtime.localtime.strftime('%Y-%m-%d %H:%M:%S %z')
+        puts "#{char_new} #{file_new}\t#{ft}"
+      end
+
+        # Loop over hunks. If a hunk overlaps with the last hunk, join them.
+        # Otherwise, print out the old one.
+      oldhunk = hunk = nil
+
+      if @format == :ed
+        real_output = output
+        output = []
+      end
+
+      diffs.each do |piece|
+      begin
+        hunk = Diff::LCS::Hunk.new(data_old, data_new, piece, @lines,
+                                   file_length_difference)
+        file_length_difference = hunk.file_length_difference
+
+        next unless oldhunk
+
+        if (@lines > 0) and hunk.overlaps?(oldhunk)
+          hunk.unshift(oldhunk)
+        else
+          output << oldhunk.diff(@format)
+        end
+      ensure
+        oldhunk = hunk
+        output << "\n"
+      end
+      end
+
+      output << oldhunk.diff(@format)
+      output << "\n"
+
+      if @format == :ed
+        output.reverse_each { |e| real_output << e.diff(:ed_finish) }
+      end
+
+      return 1
+    end
+  end
+end