dont split surrogate pairs, fixes #1

jhchen · Sep 17, 2017 · a5cecd8 · a5cecd8
1 parent de7329a
commit a5cecd8
Show file tree

Hide file tree

Showing 2 changed files with 79 additions and 0 deletions.
diff --git a/diff.js b/diff.js
@@ -82,6 +82,7 @@ function diff_main(text1, text2, cursor_pos) {
   if (cursor_pos != null) {
     diffs = fix_cursor(diffs, cursor_pos);
   }
+  diffs = fix_emoji(diffs);
   return diffs;
 };
 
@@ -671,7 +672,46 @@ function fix_cursor (diffs, cursor_pos) {
       return diffs;
     }
   }
+}
+
+/*
+ * Check diff did not split surrogate pairs.
+ * Ex. [0, '\uD83D'], [-1, '\uDC36'], [1, '\uDC2F'] -> [-1, '\uD83D\uDC36'], [1, '\uD83D\uDC2F']
+ *     '\uD83D\uDC36' === '🐶', '\uD83D\uDC2F' === '🐯'
+ *
+ * @param {Array} diffs Array of diff tuples
+ * @return {Array} Array of diff tuples
+ */
+function fix_emoji (diffs) {
+  var compact = false;
+  var starts_with_pair_end = function(str) {
+    return str.charCodeAt(0) >= 0xDC00 && str.charCodeAt(0) <= 0xDFFF;
+  }
+  var ends_with_pair_start = function(str) {
+    return str.charCodeAt(str.length-1) >= 0xD800 && str.charCodeAt(str.length-1) <= 0xDBFF;
+  }
+  for (var i = 2; i < diffs.length; i += 1) {
+    if (diffs[i-2][0] === DIFF_EQUAL && ends_with_pair_start(diffs[i-2][1]) &&
+        diffs[i-1][0] === DIFF_DELETE && starts_with_pair_end(diffs[i-1][1]) &&
+        diffs[i][0] === DIFF_INSERT && starts_with_pair_end(diffs[i][1])) {
+      compact = true;
+
+      diffs[i-1][1] = diffs[i-2][1].slice(-1) + diffs[i-1][1];
+      diffs[i][1] = diffs[i-2][1].slice(-1) + diffs[i][1];
 
+      diffs[i-2][1] = diffs[i-2][1].slice(0, -1);
+    }
+  }
+  if (!compact) {
+    return diffs;
+  }
+  var fixed_diffs = [];
+  for (var i = 0; i < diffs.length; i += 1) {
+    if (diffs[i][1].length > 0) {
+      fixed_diffs.push(diffs[i]);
+    }
+  }
+  return fixed_diffs;
 }
 
 /*

diff --git a/test.js b/test.js
@@ -49,4 +49,43 @@ for(var i = 0; i < ITERATIONS; ++i) {
   }
 }
 
+console.log('Running emoji tests');
+(function() {
+  var result = diff('🐶', '🐯');
+  var expected = [
+    [diff.DELETE, '🐶'],
+    [diff.INSERT, '🐯'],
+  ];
+  if (!_.isEqual(result, expected)) {
+    console.log(result, '!==', expected);
+    throw new Error('Emoji simple case test failed');
+  }
+})();
+
+(function() {
+  var result = diff('👨🏽', '👩🏽');
+  var expected = [
+    [diff.DELETE, '👨'],
+    [diff.INSERT, '👩'],
+    [diff.EQUAL, '🏽']
+  ];
+  if (!_.isEqual(result, expected)) {
+    console.log(result, '!==', expected);
+    throw new Error('Emoji before case test failed');
+  }
+})();
+
+(function() {
+  var result = diff('👩🏼', '👩🏽');
+  var expected = [
+    [diff.EQUAL, '👩'],
+    [diff.DELETE, '🏼'],
+    [diff.INSERT, '🏽'],
+  ];
+  if (!_.isEqual(result, expected)) {
+    console.log(result, '!==', expected);
+    throw new Error('Emoji after case test failed');
+  }
+})();
+
 console.log("Success!");