Merge branch 'pullrequest_7'

yetanotherape · Feb 11, 2017 · b00d838 · b00d838
2 parents 2b5d300 + efd1062
commit b00d838
Show file tree

Hide file tree

Showing 4 changed files with 73 additions and 44 deletions.
diff --git a/src/Diff.php b/src/Diff.php
@@ -43,6 +43,11 @@ class Diff
     const INSERT = 1;
     const EQUAL  = 0;
 
+    /**
+     * Number of characters at which line mode speedup is employed.
+     */
+    const LINEMODE_THRESOLD = 100;
+
     /**
      * @var float Number of seconds to map a diff before giving up (0 for infinity).
      */
@@ -1099,7 +1104,9 @@ protected function compute($text1, $text2, $checklines, $deadline)
             }
         }
 
-        if ($checklines && mb_strlen($text1) > 100 && mb_strlen($text2) > 100) {
+        if ($checklines
+          && mb_strlen($text1) > Diff::LINEMODE_THRESOLD
+          && mb_strlen($text2) > Diff::LINEMODE_THRESOLD) {
             return $this->lineMode($text1, $text2, $deadline);
         }
 

diff --git a/src/DiffToolkit.php b/src/DiffToolkit.php
@@ -293,44 +293,23 @@ protected function linesToCharsMunge($text, array &$lineArray, array &$lineHash)
         // explode('\n', $text) would temporarily double our memory footprint,
         // but mb_strpos() and mb_substr() work slow
         $lines = explode($delimiter, $text);
+
+        $last_line_has_delimiter = end($lines) === '';
+        if ($last_line_has_delimiter) {
+            array_pop($lines);
+        }
+
         foreach ($lines as $i => $line) {
-            if (mb_strlen($line)) {
-                if (isset($lines[$i + 1])) {
-                    $line .= $delimiter;
-                }
-                if (isset($lineHash[$line])) {
-                    $chars .= Utils::unicodeChr($lineHash[$line]);
-                } else {
-                    $lineArray[] = $line;
-                    $lineHash[$line] = count($lineArray) - 1;
-                    $chars .= Utils::unicodeChr(count($lineArray) - 1);
-                }
+            if (isset($lines[$i + 1]) || $last_line_has_delimiter) {
+                $line .= $delimiter;
             }
+            if (!isset($lineHash[$line])) {
+                $lineArray[] = $line;
+                $lineHash[$line] = count($lineArray) - 1;
+            }
+            $chars .= Utils::unicodeChr($lineHash[$line]);
         }
 
-//        // Walk the text, pulling out a substring for each line.
-//        // explode('\n', $text) would temporarily double our memory footprint.
-//        // Modifying text would create many large strings to garbage collect.
-//        $lineStart = 0;
-//        $lineEnd = -1;
-//        $textLen = mb_strlen($text);
-//        while ($lineEnd < $textLen - 1) {
-//            $lineEnd = mb_strpos($text, "\n", $lineStart);
-//            if ($lineEnd === false) {
-//                $lineEnd = $textLen - 1;
-//            }
-//            $line = mb_substr($text, $lineStart, $lineEnd + 1 - $lineStart);
-//            $lineStart = $lineEnd + 1;
-//
-//            if (isset($lineHash[$line])) {
-//                $chars .= Utils::unicodeChr($lineHash[$line]);
-//            } else {
-//                $lineArray[] = $line;
-//                $lineHash[$line] = count($lineArray) - 1;
-//                $chars .= Utils::unicodeChr(count($lineArray) - 1);
-//            }
-//        }
-
         return $chars;
     }
 

diff --git a/tests/DiffMatchPatchTest.php b/tests/DiffMatchPatchTest.php
@@ -165,13 +165,36 @@ public function testPatchMake()
         $this->assertEquals($expected, $this->dmp->patch_toText($patches));
     }
 
-    public function testPatchApply()
-    {
-        $patches = $this->dmp->patch_make("The quick brown fox jumps over the lazy dog.", "That quick brown fox jumped over a lazy dog.");
-        $this->assertEquals(
-            array("That quick red rabbit jumped over a tired tiger.", array(true, true,)),
-            $this->dmp->patch_apply($patches, "The quick red rabbit jumps over the tired tiger.")
-        );
+    protected function _testPatchApply($text1, $text2, $target = NULL, $expected = NULL) {
+      if ($target === NULL) {
+        $target = $text1;
+      }
+      if ($expected === NULL) {
+        $expected = $text2;
+      }
+
+      $patches = $this->dmp->patch_make($text1, $text2);
+      $this->assertEquals(
+        array($expected, array_map(function() { return TRUE; }, $patches)),
+        $this->dmp->patch_apply($patches, $target)
+      );
+    }
+
+    public function testPatchApply() {
+      $this->_testPatchApply(
+        "The quick brown fox jumps over the lazy dog.",
+        "That quick brown fox jumped over a lazy dog.",
+        "The quick red rabbit jumps over the tired tiger.",
+        "That quick red rabbit jumped over a tired tiger."
+      );
+    }
+
+    public function testPatchApply_2() {
+      $linemode_pad = str_pad('pad', Diff::LINEMODE_THRESOLD, 'x');
+      $this->_testPatchApply(
+        "line number one\n\nLine number three" . $linemode_pad . 'a',
+        "LINE number one\nThe second line\n\nThis is Line number three" . $linemode_pad . 'b'
+      );
     }
 
 }
diff --git a/tests/DiffTest.php b/tests/DiffTest.php
@@ -843,8 +843,28 @@ public function testMain()
         // Test the linemode speedup.
         // Must be long to pass the 100 char cutoff.
         // Simple line-mode.
-        $a = str_repeat("1234567890\n", 13);
-        $b = str_repeat("abcdefghij\n", 13);
+        $a = str_repeat("1234567890\n", ceil(Diff::LINEMODE_THRESOLD / strlen($a)));
+        $b = str_repeat("abcdefghij\n", ceil(Diff::LINEMODE_THRESOLD / strlen($b)));
+        $this->assertEquals(
+            $this->d->main($a, $b, false)->getChanges(),
+            $this->d->main($a, $b, true)->getChanges()
+        );
+
+        // Simple line-mode with multiple line breaks.
+        $a = "12345\n\n67890\n";
+        $a = str_repeat($a, ceil(Diff::LINEMODE_THRESOLD / strlen($a)));
+        $b = "abcde\n\nfghij\n";
+        $b = str_repeat($b, ceil(Diff::LINEMODE_THRESOLD / strlen($b)));
+        $this->assertEquals(
+            $this->d->main($a, $b, false)->getChanges(),
+            $this->d->main($a, $b, true)->getChanges()
+        );
+
+        // Simple line-mode with multiple line breaks but without leading line break
+        $a = "12345\n\n67890\n";
+        $a = str_repeat($a, ceil(Diff::LINEMODE_THRESOLD / strlen($a))) . '0';
+        $b = "abcde\n\nfghij\n";
+        $b = str_repeat($b, ceil(Diff::LINEMODE_THRESOLD / strlen($b))) . 'a    ';
         $this->assertEquals(
             $this->d->main($a, $b, false)->getChanges(),
             $this->d->main($a, $b, true)->getChanges()