From d7198af00a913bda8efc3f1b1b6cebef7405d190 Mon Sep 17 00:00:00 2001
From: Alexey Verkhovsky <averkhov@thoughtworks.com>
Date: Sun, 20 Feb 2005 07:46:48 +0000
Subject: [PATCH] HTMLDiff does not drop endlines anymore

---
 libraries/diff.rb      | 57 ++++++++------------------------
 test/unit/diff_test.rb | 74 ++++++++++++++++++++++++------------------
 2 files changed, 55 insertions(+), 76 deletions(-)

diff --git a/libraries/diff.rb b/libraries/diff.rb
index 12fcfaa7..45c4ee63 100644
--- a/libraries/diff.rb
+++ b/libraries/diff.rb
@@ -397,8 +397,8 @@ module HTMLDiff
   end
 
   def self.diff(a, b)
-    a, b = a.split(//), b.split(//) if a.kind_of? String and b.kind_of? String
-    a, b = html2list(a), html2list(b)
+    a = html2list(a)
+    b = html2list(b)
 
     out = Builder.new(a, b)
     s = SequenceMatcher.new(a, b)
@@ -410,37 +410,25 @@ module HTMLDiff
     out.result 
   end
   
-  def self.html2list(x, b=false)
-    mode = 'char'
+  def self.html2list(x)
+    mode = :char
     cur = ''
     out = []
     
-    x = x.split(//) if x.kind_of? String
-    
-    x.each do |c|
-      if mode == 'tag'
+    x.split('').each do |c|
+      if mode == :tag
+        cur += c
         if c == '>'
-          if b
-            cur += ']'
-          else
-            cur += c
-          end
           out.push(cur)
           cur = ''
-          mode = 'char'
-        else
-          cur += c
+          mode = :char
         end
-      elsif mode == 'char'
+      elsif mode == :char
         if c == '<'
           out.push cur
-          if b
-            cur = '['
-          else
-            cur = c
-          end
-          mode = 'tag'
-        elsif /\s/.match c
+          cur = c
+          mode = :tag
+        elsif c =~ /\s/
           out.push cur + c
           cur = ''
         else
@@ -448,28 +436,9 @@ module HTMLDiff
         end
       end
     end
-    
+
     out.push cur
-    # TODO: make something better here
-    out.each{|x| x.chomp! unless is_newline(x)}
     out.find_all { |x| x != '' }
   end
   
-  
 end
-
-if __FILE__ == $0
-
-  require 'pp'
-  # a = "<p>this is the original string</p>" # \n<p>but around the world</p>"
-  # b = "<p>this is the original </p><p>other parag</p><p>string</p>"
-  a = "<ul>\n\t<li>one</li>\n\t<li>two</li>\n</ul>"
-  b = "<ul>\n\t<li>one</li>\n\t<li>two\n\t<ul><li>abc</li></ul></li>\n</ul>"
-  puts a
-  pp HTMLDiff.html2list(a)
-  puts
-  puts b
-  pp HTMLDiff.html2list(b)
- puts
-  puts HTMLDiff.diff(a, b)
-end
\ No newline at end of file
diff --git a/test/unit/diff_test.rb b/test/unit/diff_test.rb
index b79646ff..3a0f35a5 100755
--- a/test/unit/diff_test.rb
+++ b/test/unit/diff_test.rb
@@ -7,63 +7,62 @@ include Diff
 
 class DiffTest < Test::Unit::TestCase
   def test_init
-    assert(1 == 1, "tests working")
-    assert_nothing_raised("object created") do
-      s = SequenceMatcher.new "private Thread currentThread;",
-        "private volatile Thread currentThread;",
-        proc { |x| x == ' ' }
-    end
+    assert_nothing_raised {
+      s = SequenceMatcher.new('private Thread currentThread;',
+            'private volatile Thread currentThread;') { |x| x == ' ' }
+    }
   end
   
   def test_matching_blocks
-    s = SequenceMatcher.new "abxcd", "abcd"
-    assert(s.get_matching_blocks == [[0, 0, 2], [3, 2, 2], [5, 4, 0]],
-           "get_matching_blocks works")
+    s = SequenceMatcher.new 'abxcd', 'abcd'
+    assert_equal [[0, 0, 2], [3, 2, 2], [5, 4, 0]], s.get_matching_blocks
   end
   
   def test_ratio
-    s = SequenceMatcher.new "abcd", "bcde"
-    assert(s.ratio == 0.75, "ratio works")
-    assert(s.quick_ratio == 0.75, "quick_ratio works")
-    assert(s.real_quick_ratio == 1.0, "real_quick_ratio works")
+    s = SequenceMatcher.new 'abcd', 'bcde'
+    assert_equal 0.75, s.ratio, 0.001
+    assert_equal 0.75, s.quick_ratio, 0.001
+    assert_equal 1.0, s.real_quick_ratio, 0.001
   end
   
   def test_longest_match
-    s = SequenceMatcher.new(" abcd", "abcd abcd")
-    assert(s.find_longest_match(0, 5, 0, 9) == [0, 4, 5], 
-           "find_longest_match works")
-    s = SequenceMatcher.new()
+    s = SequenceMatcher.new(' abcd', 'abcd abcd')
+    assert_equal [0, 4, 5], s.find_longest_match(0, 5, 0, 9)
   end
   
   def test_opcodes
-    s = SequenceMatcher.new("qabxcd", "abycdf")
-    assert(s.get_opcodes == [
-             [:delete, 0, 1, 0, 0],
-             [:equal, 1, 3, 0, 2],
-             [:replace, 3, 4, 2, 3],
-             [:equal, 4, 6, 3, 5],
-             [:insert, 6, 6, 5, 6]], "get_opcodes works")
+    s = SequenceMatcher.new('qabxcd', 'abycdf')
+    assert_equal( 
+      [
+        [:delete, 0, 1, 0, 0],
+        [:equal, 1, 3, 0, 2],
+        [:replace, 3, 4, 2, 3],
+        [:equal, 4, 6, 3, 5],
+        [:insert, 6, 6, 5, 6]
+      ],
+      s.get_opcodes)
   end
 
 
   def test_count_leading
-    assert(Diff.count_leading('   abc', ' ') == 3, 
-           "count_leading works")
+    assert_equal 3, Diff.count_leading('   abc', ' ')
   end
 
   def test_html2list
     a = "here is the original text"
-    #p HTMLDiff.html2list(a)
+    assert_equal(
+        ['here ', 'is ', 'the ', 'original ', 'text'],
+        HTMLDiff.html2list(a))
   end
 
   def test_html_diff
-    a = "this was the original string"
-    b = "this is the super string"
-    assert_equal 'this <del class="diffmod">was </del>' + 
+    a = 'this was the original string'
+    b = 'this is the super string'
+    assert_equal('this <del class="diffmod">was </del>' + 
            '<ins class="diffmod">is </ins>the ' +
            '<del class="diffmod">original </del>' + 
            '<ins class="diffmod">super </ins>string',
-           HTMLDiff.diff(a, b)
+           HTMLDiff.diff(a, b))
   end
   
   def test_html_diff_with_multiple_paragraphs
@@ -79,4 +78,15 @@ class DiffTest < Test::Unit::TestCase
       HTMLDiff.diff(a, b)
     )
   end
-end
\ No newline at end of file
+  
+  # FIXME this test fails (ticket #67, http://dev.instiki.org/ticket/67)
+  def test_html_diff_preserves_endlines_in_pre
+    a = "<pre>\na\nb\nc\n</pre>"
+    b = ''
+
+    assert_equal(
+        "<pre>\n<del class=\"diffdel\">a\nb\nc\n</del></pre>",
+        HTMLDiff.diff(a, b))
+  end
+  
+end