diff-lcs 1.4.1 → 1.5.0

data/Manifest.txt

@@ -6,7 +6,6 @@ License.md
 Manifest.txt
 README.rdoc
 Rakefile
-autotest/discover.rb
 bin/htmldiff
 bin/ldiff
 docs/COPYING.txt
@@ -34,6 +33,21 @@ spec/fixtures/ldiff/output.diff-c
 spec/fixtures/ldiff/output.diff-e
 spec/fixtures/ldiff/output.diff-f
 spec/fixtures/ldiff/output.diff-u
+spec/fixtures/ldiff/output.diff.chef
+spec/fixtures/ldiff/output.diff.chef-c
+spec/fixtures/ldiff/output.diff.chef-e
+spec/fixtures/ldiff/output.diff.chef-f
+spec/fixtures/ldiff/output.diff.chef-u
+spec/fixtures/ldiff/output.diff.chef2
+spec/fixtures/ldiff/output.diff.chef2-c
+spec/fixtures/ldiff/output.diff.chef2-d
+spec/fixtures/ldiff/output.diff.chef2-e
+spec/fixtures/ldiff/output.diff.chef2-f
+spec/fixtures/ldiff/output.diff.chef2-u
+spec/fixtures/new-chef
+spec/fixtures/new-chef2
+spec/fixtures/old-chef
+spec/fixtures/old-chef2
 spec/hunk_spec.rb
 spec/issues_spec.rb
 spec/lcs_spec.rb

data/README.rdoc

@@ -12,14 +12,15 @@ Diff::LCS computes the difference between two Enumerable sequences using the
 McIlroy-Hunt longest common subsequence (LCS) algorithm. It includes utilities
 to create a simple HTML diff output format and a standard diff-like tool.
 
-This is release 1.4, providing a simple extension that allows for
-Diff::LCS::Change objects to be treated implicitly as arrays. Ruby versions
-below 2.5 are soft-deprecated.
-
-This means that older versions are no longer part of the CI test suite. If any
-changes have been introduced that break those versions, bug reports and patches
-will be accepted, but it will be up to the reporter to verify any fixes prior
-to release. A future release will completely break compatibility.
+This is release 1.4.3, providing a simple extension that allows for
+Diff::LCS::Change objects to be treated implicitly as arrays and fixes a
+number of formatting issues.
+
+Ruby versions below 2.5 are soft-deprecated, which means that older versions
+are no longer part of the CI test suite. If any changes have been introduced
+that break those versions, bug reports and patches will be accepted, but it
+will be up to the reporter to verify any fixes prior to release. The next
+major release will completely break compatibility.
 
 == Synopsis
 

data/Rakefile

@@ -2,14 +2,68 @@
 
 require 'rubygems'
 require 'rspec'
+require 'rspec/core/rake_task'
 require 'hoe'
 
+# This is required until https://github.com/seattlerb/hoe/issues/112 is fixed
+class Hoe
+  def with_config
+    config = Hoe::DEFAULT_CONFIG
+
+    rc = File.expand_path("~/.hoerc")
+    homeconfig = load_config(rc)
+    config = config.merge(homeconfig)
+
+    localconfig = load_config(File.expand_path(File.join(Dir.pwd, ".hoerc")))
+    config = config.merge(localconfig)
+
+    yield config, rc
+  end
+
+  def load_config(name)
+    File.exist?(name) ? safe_load_yaml(name) : {}
+  end
+
+  def safe_load_yaml(name)
+    return safe_load_yaml_file(name) if YAML.respond_to?(:safe_load_file)
+
+    data = IO.binread(name)
+    YAML.safe_load(data, permitted_classes: [Regexp])
+  rescue
+    YAML.safe_load(data, [Regexp])
+  end
+
+  def safe_load_yaml_file(name)
+    YAML.safe_load_file(name, permitted_classes: [Regexp])
+  rescue
+    YAML.safe_load_file(name, [Regexp])
+  end
+end
+
 Hoe.plugin :bundler
 Hoe.plugin :doofus
-Hoe.plugin :email unless ENV['CI'] or ENV['TRAVIS']
 Hoe.plugin :gemspec2
 Hoe.plugin :git
-Hoe.plugin :travis
+
+if RUBY_VERSION < '1.9'
+  class Array #:nodoc:
+    def to_h
+      Hash[*flatten(1)]
+    end
+  end
+
+  class Gem::Specification #:nodoc:
+    def metadata=(*); end
+
+    def default_value(*); end
+  end
+
+  class Object #:nodoc:
+    def caller_locations(*)
+      []
+    end
+  end
+end
 
 _spec = Hoe.spec 'diff-lcs' do
   developer('Austin Ziegler', 'halostatue@gmail.com')
@@ -26,9 +80,20 @@ _spec = Hoe.spec 'diff-lcs' do
   extra_dev_deps << ['hoe-rubygems', '~> 1.0']
   extra_dev_deps << ['rspec', '>= 2.0', '< 4']
   extra_dev_deps << ['rake', '>= 10.0', '< 14']
-  extra_dev_deps << ['rdoc', '>= 0']
+  extra_dev_deps << ['rdoc', '>= 6.3.1', '< 7']
+end
+
+desc "Run all specifications"
+RSpec::Core::RakeTask.new(:spec) do |t|
+  rspec_dirs = %w(spec lib).join(":")
+  t.rspec_opts = ["-I#{rspec_dirs}"]
 end
 
+Rake::Task["spec"].actions.uniq! { |a| a.source_location }
+
+task :default => :spec unless Rake::Task["default"].prereqs.include?("spec")
+task :test => :spec unless Rake::Task["test"].prereqs.include?("spec")
+
 if RUBY_VERSION >= '2.0' && RUBY_ENGINE == 'ruby'
   namespace :spec do
     desc "Runs test coverage. Only works Ruby 2.0+ and assumes 'simplecov' is installed."
@@ -38,3 +103,19 @@ if RUBY_VERSION >= '2.0' && RUBY_ENGINE == 'ruby'
     end
   end
 end
+
+task :ruby18 do
+  puts <<-MESSAGE
+You are starting a barebones Ruby 1.8 docker environment. You will need to
+do the following:
+
+- mv Gemfile.lock{,.v2}
+- gem install bundler --version 1.17.2 --no-ri --no-rdoc
+- ruby -S bundle
+- rake
+
+Don't forget to restore your Gemfile.lock after testing.
+
+  MESSAGE
+  sh "docker run -it --rm -v #{Dir.pwd}:/root/diff-lcs bellbind/docker-ruby18-rails2 bash -l"
+end

data/lib/diff/lcs/backports.rb

@@ -3,7 +3,7 @@
 unless 0.respond_to?(:positive?)
   class Fixnum # rubocop:disable Lint/UnifiedInteger, Style/Documentation
     def positive?
-      self > 0 # rubocop:disable Styel/NumericPredicate
+      self > 0 # rubocop:disable Style/NumericPredicate
     end
   end
 end

data/lib/diff/lcs/block.rb

@@ -19,7 +19,7 @@ class Diff::LCS::Block
   end
 
   def diff_size
-    (@insert.size - @remove.size).abs
+    @insert.size - @remove.size
   end
 
   def op

data/lib/diff/lcs/hunk.rb

@@ -2,30 +2,42 @@
 
 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).
+# 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/ldiff).
 class Diff::LCS::Hunk
-  # Create a hunk using references to both the old and new data, as well as
-  # the piece of data.
+  OLD_DIFF_OP_ACTION = { '+' => 'a', '-' => 'd', '!' => 'c' }.freeze #:nodoc:
+  ED_DIFF_OP_ACTION = { '+' => 'a', '-' => 'd', '!' => 'c' }.freeze #:nodoc:
+
+  private_constant :OLD_DIFF_OP_ACTION, :ED_DIFF_OP_ACTION if respond_to?(:private_constant)
+
+  # 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, flag_context, file_length_difference)
     # At first, a hunk will have just one Block in it
     @blocks = [Diff::LCS::Block.new(piece)]
+
+    if @blocks[0].remove.empty? && @blocks[0].insert.empty?
+      fail "Cannot build a hunk from #{piece.inspect}; has no add or remove actions"
+    end
+
     if String.method_defined?(:encoding)
       @preferred_data_encoding = data_old.fetch(0, data_new.fetch(0, '')).encoding
     end
+
     @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
-    @max_diff_size = @blocks.lazy.map { |e| e.diff_size }.max
+    @max_diff_size = @blocks.map { |e| e.diff_size.abs }.max
+
 
     # 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.
+    # 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
@@ -55,23 +67,26 @@ class Diff::LCS::Hunk
 
   # Change the "start" and "end" fields to note that context should be added
   # to this hunk.
-  attr_accessor :flag_context
+  attr_accessor :flag_context # rubocop:disable Layout/EmptyLinesAroundAttributeAccessor
   undef :flag_context=
   def flag_context=(context) #:nodoc: # rubocop:disable Lint/DuplicateMethods
     return if context.nil? or context.zero?
 
     add_start = context > @start_old ? @start_old : context
+
     @start_old -= add_start
     @start_new -= add_start
 
+    old_size = @data_old.size
+
     add_end =
-      if (@end_old + context) > @data_old.size
-        @data_old.size - @end_old
+      if (@end_old + context) > old_size
+        old_size - @end_old
       else
         context
       end
 
-    add_end = @max_diff_size if add_end > @max_diff_size
+    add_end = @max_diff_size if add_end >= old_size
 
     @end_old += add_end
     @end_new += add_end
@@ -98,18 +113,18 @@ class Diff::LCS::Hunk
   end
 
   # Returns a diff string based on a format.
-  def diff(format)
+  def diff(format, last = false)
     case format
     when :old
-      old_diff
+      old_diff(last)
     when :unified
-      unified_diff
+      unified_diff(last)
     when :context
-      context_diff
+      context_diff(last)
     when :ed
       self
     when :reverse_ed, :ed_finish
-      ed_diff(format)
+      ed_diff(format, last)
     else
       fail "Unknown diff format #{format}."
     end
@@ -117,35 +132,34 @@ class Diff::LCS::Hunk
 
   # 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
+  def old_diff(_last = false)
     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 = encode("#{context_range(:old)}#{op_act[block.op]}#{context_range(:new)}\n")
+    s = encode("#{context_range(:old, ',')}#{OLD_DIFF_OP_ACTION[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.
     unless block.remove.empty?
-      @data_old[@start_old..@end_old].each { |e| s << encode('< ') + e + encode("\n") }
+      @data_old[@start_old..@end_old].each { |e| s << encode('< ') + e.chomp + encode("\n") }
     end
 
     s << encode("---\n") if block.op == '!'
 
     unless block.insert.empty?
-      @data_new[@start_new..@end_new].each { |e| s << encode('> ') + e + encode("\n") }
+      @data_new[@start_new..@end_new].each { |e| s << encode('> ') + e.chomp + encode("\n") }
     end
 
     s
   end
   private :old_diff
 
-  def unified_diff
+  def unified_diff(last = false)
     # Calculate item number range.
-    s = encode("@@ -#{unified_range(:old)} +#{unified_range(:new)} @@\n")
+    s = encode("@@ -#{unified_range(:old, last)} +#{unified_range(:new, last)} @@\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
@@ -158,7 +172,14 @@ class Diff::LCS::Hunk
     # 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].map { |e| e.insert(0, encode(' ')) }
+    outlist = @data_old[lo..hi].map { |e| String.new("#{encode(' ')}#{e.chomp}") }
+
+    last_block = blocks[-1]
+
+    if last
+      old_missing_newline = missing_last_newline?(@data_old)
+      new_missing_newline = missing_last_newline?(@data_new)
+    end
 
     @blocks.each do |block|
       block.remove.each do |item|
@@ -167,67 +188,100 @@ class Diff::LCS::Hunk
         outlist[offset][0, 1] = encode(op)
         num_removed += 1
       end
+
+      if last && block == last_block && old_missing_newline && !new_missing_newline
+        outlist << encode('\')
+        num_removed += 1
+      end
+
       block.insert.each do |item|
         op     = item.action.to_s # +
         offset = item.position - @start_new + num_removed
-        outlist[offset, 0] = encode(op) + @data_new[item.position]
+        outlist[offset, 0] = encode(op) + @data_new[item.position].chomp
         num_added += 1
       end
     end
 
+    outlist << encode('\') if last && new_missing_newline
+
     s << outlist.join(encode("\n"))
+
+    s
   end
   private :unified_diff
 
-  def context_diff
+  def context_diff(last = false)
     s = encode("***************\n")
-    s << encode("*** #{context_range(:old)} ****\n")
-    r = context_range(:new)
+    s << encode("*** #{context_range(:old, ',', last)} ****\n")
+    r = context_range(:new, ',', last)
+
+    if last
+      old_missing_newline = missing_last_newline?(@data_old)
+      new_missing_newline = missing_last_newline?(@data_new)
+    end
 
     # 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.reject { |e| e.remove.empty? }
-    if removes
-      outlist = @data_old[lo..hi].map { |e| e.insert(0, encode(' ')) }
+
+    unless removes.empty?
+      outlist = @data_old[lo..hi].map { |e| String.new("#{encode('  ')}#{e.chomp}") }
+
+      last_block = removes[-1]
 
       removes.each do |block|
         block.remove.each do |item|
-          outlist[item.position - lo].insert(0, encode(block.op)) # - or !
+          outlist[item.position - lo][0, 1] = encode(block.op) # - or !
+        end
+
+        if last && block == last_block && old_missing_newline
+          outlist << encode('\')
         end
       end
-      s << outlist.join("\n")
+
+      s << outlist.join(encode("\n")) << encode("\n")
     end
 
-    s << encode("\n--- #{r} ----\n")
+    s << encode("--- #{r} ----\n")
     lo, hi = @start_new, @end_new
     inserts = @blocks.reject { |e| e.insert.empty? }
-    if inserts
-      outlist = @data_new[lo..hi].collect { |e| e.insert(0, encode(' ')) }
+
+    unless inserts.empty?
+      outlist = @data_new[lo..hi].map { |e| String.new("#{encode('  ')}#{e.chomp}") }
+
+      last_block = inserts[-1]
+
       inserts.each do |block|
         block.insert.each do |item|
-          outlist[item.position - lo].insert(0, encode(block.op)) # - or !
+          outlist[item.position - lo][0, 1] = encode(block.op) # + or !
+        end
+
+        if last && block == last_block && new_missing_newline
+          outlist << encode('\')
         end
       end
-      s << outlist.join("\n")
+      s << outlist.join(encode("\n"))
     end
+
     s
   end
   private :context_diff
 
-  def ed_diff(format)
-    op_act = { '+' => 'a', '-' => 'd', '!' => 'c' }
+  def ed_diff(format, _last = false)
     warn 'Expecting only one block in an old diff hunk!' if @blocks.size > 1
 
     s =
       if format == :reverse_ed
-        encode("#{op_act[@blocks[0].op]}#{context_range(:old)}\n")
+        encode("#{ED_DIFF_OP_ACTION[@blocks[0].op]}#{context_range(:old, ',')}\n")
       else
-        encode("#{context_range(:old, ' ')}#{op_act[@blocks[0].op]}\n")
+        encode("#{context_range(:old, ' ')}#{ED_DIFF_OP_ACTION[@blocks[0].op]}\n")
       end
 
     unless @blocks[0].insert.empty?
-      @data_new[@start_new..@end_new].each do |e| s << e + encode("\n") end
+      @data_new[@start_new..@end_new].each do |e|
+        s << e.chomp + encode("\n")
+      end
       s << encode(".\n")
     end
     s
@@ -236,7 +290,7 @@ class Diff::LCS::Hunk
 
   # 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, op = ',') # rubocop:disable Naming/UncommunicativeMethodParamName
+  def context_range(mode, op, last = false)
     case mode
     when :old
       s, e = (@start_old + 1), (@end_old + 1)
@@ -244,6 +298,9 @@ class Diff::LCS::Hunk
       s, e = (@start_new + 1), (@end_new + 1)
     end
 
+    e -= 1 if last
+    e = 1 if e.zero?
+
     s < e ? "#{s}#{op}#{e}" : e.to_s
   end
   private :context_range
@@ -251,7 +308,7 @@ class Diff::LCS::Hunk
   # 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)
+  def unified_range(mode, last)
     case mode
     when :old
       s, e = (@start_old + 1), (@end_old + 1)
@@ -259,12 +316,25 @@ class Diff::LCS::Hunk
       s, e = (@start_new + 1), (@end_new + 1)
     end
 
-    length = e - s + 1
+    length = e - s + (last ? 0 : 1)
+
     first = length < 2 ? e : s # "strange, but correct"
-    length == 1 ? first.to_s : "#{first},#{length}"
+    length <= 1 ? first.to_s : "#{first},#{length}"
   end
   private :unified_range
 
+  def missing_last_newline?(data)
+    newline = encode("\n")
+
+    if data[-2]
+      data[-2].end_with?(newline) && !data[-1].end_with?(newline)
+    elsif data[-1]
+      !data[-1].end_with?(newline)
+    else
+      true
+    end
+  end
+
   if String.method_defined?(:encoding)
     def encode(literal, target_encoding = @preferred_data_encoding)
       literal.encode target_encoding

data/lib/diff/lcs/internals.rb

@@ -44,13 +44,12 @@ class << Diff::LCS::Internals
     b_finish = b.size - 1
     vector = []
 
-    # Prune off any common elements at the beginning...
+    # Collect any common elements at the beginning...
     while (a_start <= a_finish) and (b_start <= b_finish) and (a[a_start] == b[b_start])
       vector[a_start] = b_start
       a_start += 1
       b_start += 1
     end
-    b_start = a_start
 
     # Now the end...
     while (a_start <= a_finish) and (b_start <= b_finish) and (a[a_finish] == b[b_finish])
@@ -60,6 +59,7 @@ class << Diff::LCS::Internals
     end
 
     # Now, compute the equivalence classes of positions of elements.
+    # An explanation for how this works: https://codeforces.com/topic/92191
     b_matches = position_hash(b, b_start..b_finish)
 
     thresh = []
@@ -71,6 +71,10 @@ class << Diff::LCS::Internals
       bm = b_matches[ai]
       k = nil
       bm.reverse_each do |j|
+        # Although the threshold check is not mandatory for this to work,
+        # it may have an optimization purpose
+        # An attempt to remove it: https://github.com/halostatue/diff-lcs/pull/72
+        # Why it is reintroduced: https://github.com/halostatue/diff-lcs/issues/78
         if k and (thresh[k] > j) and (thresh[k - 1] < j)
           thresh[k] = j
         else
@@ -253,7 +257,7 @@ enumerable as either source or destination value."
     end
 
     # Binary search for the insertion point
-    last_index ||= enum.size
+    last_index ||= enum.size - 1
     first_index = 0
     while first_index <= last_index
       i = (first_index + last_index) >> 1

data/lib/diff/lcs/ldiff.rb

@@ -98,24 +98,19 @@ class << Diff::LCS::Ldiff
     # 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].scan(/\0/).empty?
-        new_txt = data_new[0, 4096].scan(/\0/).empty?
-        @binary = !old_txt or !new_txt
-      end
+    data_old = IO.read(file_old)
+    data_new = IO.read(file_new)
+
+    # Test binary status
+    if @binary.nil?
+      old_txt = data_old[0, 4096].scan(/\0/).empty?
+      new_txt = data_new[0, 4096].scan(/\0/).empty?
+      @binary = !old_txt || !new_txt
+    end
 
-      unless @binary
-        data_old = data_old.split($/).map { |e| e.chomp }
-        data_new = data_new.split($/).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 }
+    unless @binary
+      data_old = data_old.lines.to_a
+      data_new = data_new.lines.to_a
     end
 
     # diff yields lots of pieces, each of which is basically a Block object
@@ -150,20 +145,21 @@ class << Diff::LCS::Ldiff
     end
 
     diffs.each do |piece|
-      begin
+      begin # rubocop:disable Style/RedundantBegin
         hunk = Diff::LCS::Hunk.new(data_old, data_new, piece, @lines, file_length_difference)
         file_length_difference = hunk.file_length_difference
 
         next unless oldhunk
         next if @lines.positive? and hunk.merge(oldhunk)
 
-        output << oldhunk.diff(@format) << "\n"
+        output << oldhunk.diff(@format)
+        output << "\n" if @format == :unified
       ensure
         oldhunk = hunk
       end
     end
 
-    last = oldhunk.diff(@format)
+    last = oldhunk.diff(@format, true)
     last << "\n" if last.respond_to?(:end_with?) && !last.end_with?("\n")
 
     output << last