diff-lcs 1.3 → 1.6.2

data/lib/diff/lcs/hunk.rb

@@ -1,30 +1,44 @@
-# -*- ruby encoding: utf-8 -*-
+# frozen_string_literal: true
 
-require 'diff/lcs/block'
+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) ]
+    @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
+      @preferred_data_encoding = data_old.fetch(0) { data_new.fetch(0) { "" } }.encoding
     end
+
     @data_old = data_old
     @data_new = data_new
+    @old_empty = data_old.empty? || (data_old.size == 1 && data_old[0].empty?)
+    @new_empty = data_new.empty? || (data_new.size == 1 && data_new[0].empty?)
 
     before = after = file_length_difference
     after += @blocks[0].diff_size
     @file_length_difference = after # The caller must get this manually
+    @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
@@ -41,8 +55,8 @@ class Diff::LCS::Hunk
 
     @start_old = a1 || (b1 - before)
     @start_new = b1 || (a1 + before)
-    @end_old   = a2 || (b2 - after)
-    @end_new   = b2 || (a2 + after)
+    @end_old = a2 || (b2 - after)
+    @end_new = b2 || (a2 + after)
 
     self.flag_context = flag_context
   end
@@ -55,19 +69,24 @@ class Diff::LCS::Hunk
   # Change the "start" and "end" fields to note that context should be added
   # to this hunk.
   attr_accessor :flag_context
-  undef :flag_context=;
-  def flag_context=(context) #:nodoc:
-    return if context.nil? or context.zero?
+  undef :flag_context=
+  def flag_context=(context) # :nodoc: # standard:disable Lint/DuplicateMethods
+    return if context.nil? || 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
+    old_size = @data_old.size
+
+    add_end =
+      if (@end_old + context) >= old_size
+        old_size - @end_old - 1
+      else
+        context
+      end
+
     @end_old += add_end
     @end_new += add_end
   end
@@ -76,13 +95,11 @@ class Diff::LCS::Hunk
   # a truthy value so that if there is no overlap, you can know the merge
   # was skipped.
   def merge(hunk)
-    if overlaps?(hunk)
-      @start_old = hunk.start_old
-      @start_new = hunk.start_new
-      blocks.unshift(*hunk.blocks)
-    else
-      nil
-    end
+    return unless overlaps?(hunk)
+
+    @start_old = hunk.start_old
+    @start_new = hunk.start_new
+    blocks.unshift(*hunk.blocks)
   end
   alias_method :unshift, :merge
 
@@ -95,45 +112,59 @@ 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
-      raise "Unknown diff format #{format}."
+      fail "Unknown diff format #{format}."
     end
   end
 
   # 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]
 
+    if last
+      old_missing_newline = !@old_empty && missing_last_newline?(@data_old)
+      new_missing_newline = !@new_empty && missing_last_newline?(@data_new)
+    end
+
     # 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.
-    @data_old[@start_old .. @end_old].each { |e| s << encode("< ") + e + encode("\n") } unless block.remove.empty?
+    unless block.remove.empty?
+      @data_old[@start_old..@end_old].each { |e| s << encode("< ") + e.chomp + encode("\n") }
+    end
+
+    s << encode("\\n") if old_missing_newline && !new_missing_newline
     s << encode("---\n") if block.op == "!"
-    @data_new[@start_new .. @end_new].each { |e| s << encode("> ") + e + encode("\n") } unless block.insert.empty?
+
+    unless block.insert.empty?
+      @data_new[@start_new..@end_new].each { |e| s << encode("> ") + e.chomp + encode("\n") }
+    end
+
+    s << encode("\\n") if new_missing_newline && !old_missing_newline
+
     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")
 
@@ -148,75 +179,132 @@ 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(' ')) }
+    # standard:disable Performance/UnfreezeString
+    outlist = @data_old[lo..hi].map { |e| String.new("#{encode(" ")}#{e.chomp}") }
+    # standard:enable Performance/UnfreezeString
+
+    last_block = blocks[-1]
+
+    if last
+      old_missing_newline = !@old_empty && missing_last_newline?(@data_old)
+      new_missing_newline = !@new_empty && missing_last_newline?(@data_new)
+    end
 
     @blocks.each do |block|
       block.remove.each do |item|
-        op     = item.action.to_s # -
+        op = item.action.to_s # -
         offset = item.position - lo + num_added
         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 # +
+        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, ",")} ****\n")
+    r = context_range(:new, ",")
+
+    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.select { |e| not e.remove.empty? }
-    if removes
-      outlist = @data_old[lo .. hi].map { |e| e.insert(0, encode(' ')) }
+    removes = @blocks.reject { |e| e.remove.empty? }
+
+    unless removes.empty?
+      # standard:disable Performance/UnfreezeString
+      outlist = @data_old[lo..hi].map { |e| String.new("#{encode("  ")}#{e.chomp}") }
+      # standard:enable Performance/UnfreezeString
+
+      last_block = removes[-1]
 
       removes.each do |block|
         block.remove.each do |item|
           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.select { |e| not e.insert.empty? }
-    if inserts
-      outlist = @data_new[lo .. hi].collect { |e| e.insert(0, encode(' ')) }
+    inserts = @blocks.reject { |e| e.insert.empty? }
+
+    unless inserts.empty?
+      # standard:disable Performance/UnfreezeString
+      outlist = @data_new[lo..hi].map { |e| String.new("#{encode("  ")}#{e.chomp}") }
+      # standard:enable Performance/UnfreezeString
+
+      last_block = inserts[-1]
+
       inserts.each do |block|
         block.insert.each do |item|
           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)
     warn "Expecting only one block in an old diff hunk!" if @blocks.size > 1
-
-    if format == :reverse_ed
-      s = encode("#{op_act[@blocks[0].op]}#{context_range(:old)}\n")
-    else
-      s = encode("#{context_range(:old, ' ')}#{op_act[@blocks[0].op]}\n")
+    if last
+      # ed script doesn't support well incomplete lines
+      warn "<old_file>: No newline at end of file\n" if !@old_empty && missing_last_newline?(@data_old)
+      warn "<new_file>: No newline at end of file\n" if !@new_empty && missing_last_newline?(@data_new)
+
+      if @blocks[0].op == "!"
+        return +"" if @blocks[0].changes[0].element == @blocks[0].changes[1].element + "\n"
+        return +"" if @blocks[0].changes[0].element + "\n" == @blocks[0].changes[1].element
+      end
     end
 
+    s =
+      if format == :reverse_ed
+        encode("#{ED_DIFF_OP_ACTION[@blocks[0].op]}#{context_range(:old, " ")}\n")
+      else
+        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 { |e| s << e + encode("\n") }
+      @data_new[@start_new..@end_new].each do |e|
+        s << e.chomp + encode("\n")
+      end
       s << encode(".\n")
     end
     s
@@ -225,7 +313,7 @@ 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 = ',')
+  def context_range(mode, op)
     case mode
     when :old
       s, e = (@start_old + 1), (@end_old + 1)
@@ -233,7 +321,7 @@ class Diff::LCS::Hunk
       s, e = (@start_new + 1), (@end_new + 1)
     end
 
-    (s < e) ? "#{s}#{op}#{e}" : "#{e}"
+    (s < e) ? "#{s}#{op}#{e}" : e.to_s
   end
   private :context_range
 
@@ -243,17 +331,31 @@ class Diff::LCS::Hunk
   def unified_range(mode)
     case mode
     when :old
+      return "0,0" if @old_empty
       s, e = (@start_old + 1), (@end_old + 1)
     when :new
+      return "0,0" if @new_empty
       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}"
+
+    (length <= 1) ? e.to_s : "#{s},#{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
@@ -263,10 +365,11 @@ class Diff::LCS::Hunk
       args.map { |arg| arg.encode(string.encoding) }
     end
   else
-    def encode(literal, target_encoding = nil)
+    def encode(literal, _target_encoding = nil)
       literal
     end
-    def encode_as(string, *args)
+
+    def encode_as(_string, *args)
       args
     end
   end