Refactor Java version to use cleanupSplitSurrogates

google · May 10, 2023 · d681ef6 · d681ef6
1 parent 191b698
commit d681ef6
Showing 1 changed file with 41 additions and 24 deletions.
diff --git a/java/src/name/fraser/neil/plaintext/diff_match_patch.java b/java/src/name/fraser/neil/plaintext/diff_match_patch.java
@@ -1294,6 +1294,46 @@ public void diff_cleanupMerge(LinkedList<Diff> diffs) {
     }
   }
 
+  /**
+   * Rearrange diff boudnaries that split Unicode surrogate pairs.
+   * @param diffs Linked list of diff objects
+   */
+  public void diff_cleanupSplitSurrogates(List<Diff> diffs) {
+    char lastEnd = 0;
+    boolean isFirst = true;
+    HashSet<Diff> toRemove = new HashSet<Diff>();
+
+    for (Diff aDiff : diffs) {
+      if (aDiff.text.isEmpty()) {
+        toRemove.add(aDiff);
+        continue;
+      }
+
+      char thisTop = aDiff.text.charAt(0);
+      char thisEnd = aDiff.text.charAt(aDiff.text.length() - 1);
+
+      if (Character.isHighSurrogate(thisEnd)) {
+        lastEnd = thisEnd;
+        aDiff.text = aDiff.text.substring(0, aDiff.text.length() - 1);
+      }
+
+      if (!isFirst && Character.isHighSurrogate(lastEnd) && Character.isLowSurrogate(thisTop)) {
+        aDiff.text = lastEnd + aDiff.text;
+      }
+
+      isFirst = false;
+
+      if ( aDiff.text.isEmpty() ) {
+        toRemove.add(aDiff);
+        continue;
+      }
+    }
+
+    for (Diff aDiff : toRemove) {
+      diffs.remove(aDiff);
+    }
+  }
+
   /**
    * loc is a location in text1, compute and return the equivalent location in
    * text2.
@@ -1430,31 +1470,8 @@ public int diff_levenshtein(List<Diff> diffs) {
    */
   public String diff_toDelta(List<Diff> diffs) {
     StringBuilder text = new StringBuilder();
-    char lastEnd = 0;
-    boolean isFirst = true;
+    this.diff_cleanupSplitSurrogates(diffs);
     for (Diff aDiff : diffs) {
-      if (aDiff.text.isEmpty()) {
-        continue;
-      }
-
-      char thisTop = aDiff.text.charAt(0);
-      char thisEnd = aDiff.text.charAt(aDiff.text.length() - 1);
-
-      if (Character.isHighSurrogate(thisEnd)) {
-        lastEnd = thisEnd;
-        aDiff.text = aDiff.text.substring(0, aDiff.text.length() - 1);
-      }
-
-      if (! isFirst && Character.isHighSurrogate(lastEnd) && Character.isLowSurrogate(thisTop)) {
-        aDiff.text = lastEnd + aDiff.text;
-      }
-
-      isFirst = false;
-
-      if ( aDiff.text.isEmpty() ) {
-        continue;
-      }
-
       switch (aDiff.operation) {
       case INSERT:
         try {