[CODEC-331]

org.apache.commons.codec.language.bm.Rule.parsePhonemeExpr(String) adds
duplicate empty phoneme when input ends with |

Author: garydgregory

src/changes/changes.xml

@@ -56,6 +56,7 @@ The <action> type attribute can be add,update,fix,remove.
       <action type="fix" dev="ggregory" due-to="Gary Gregory">Remove redundant checks for whitespace in DaitchMokotoffSoundex.soundex(String, boolean).</action>
       <action type="fix" dev="ggregory" due-to="Sebastian Baunsgaard">Javadoc typo in Base16.java #380.</action>
       <action type="fix" dev="ggregory" due-to="Gary Gregory">Deprecate unused constant org.apache.commons.codec.language.bm.Rule.ALL.</action>
+      <action type="fix" dev="ggregory" issue="CODEC-331" due-to="IlikeCode, Gary Gregory">org.apache.commons.codec.language.bm.Rule.parsePhonemeExpr(String) adds duplicate empty phoneme when input ends with |.</action>
       <!-- ADD -->
       <action type="add" dev="ggregory" due-to="Gary Gregory">Add HmacUtils.hmac(Path).</action>      
       <action type="add" dev="ggregory" due-to="Gary Gregory">Add HmacUtils.hmacHex(Path).</action>      

src/main/java/org/apache/commons/codec/language/bm/Rule.java

@@ -80,6 +80,8 @@
  */
 public class Rule {
 
+    private static final String PIPE = "|";
+
     /**
      * A phoneme.
      */
@@ -426,18 +428,19 @@ private static Phoneme parsePhoneme(final String ph) {
         return new Phoneme(ph, Languages.ANY_LANGUAGE);
     }
 
-    private static PhonemeExpr parsePhonemeExpr(final String ph) {
+    static PhonemeExpr parsePhonemeExpr(final String ph) {
         if (ph.startsWith("(")) {
             // we have a bracketed list of options
             if (!ph.endsWith(")")) {
-                throw new IllegalArgumentException("Phoneme starts with '(' so must end with ')'");
+                throw new IllegalArgumentException("Phoneme starting with '(' must end with ')'");
             }
             final List<Phoneme> phs = new ArrayList<>();
             final String body = ph.substring(1, ph.length() - 1);
-            for (final String part : AROUND_PIPE.split(body)) {
+            final String[] split = AROUND_PIPE.split(body);
+            for (final String part : split) {
                 phs.add(parsePhoneme(part));
             }
-            if (body.startsWith("|") || body.endsWith("|")) {
+            if (split.length > 1 && split[0].length() != 0 && body.startsWith(PIPE) || split[split.length - 1].length() != 0 && body.endsWith(PIPE)) {
                 phs.add(new Phoneme("", Languages.ANY_LANGUAGE));
             }
             return new PhonemeList(phs);

src/test/java/org/apache/commons/codec/language/bm/RuleTest.java

@@ -63,6 +63,15 @@ void testPhonemeComparedToSelfIsZero() {
         }
     }
 
+    @Test
+    void testParsePhonemeExprLang311() {
+        assertEquals(1, Rule.parsePhonemeExpr("()").size());
+        assertEquals(1, Rule.parsePhonemeExpr("(())").size());
+        assertEquals(2, Rule.parsePhonemeExpr("(()|)").size());
+        assertEquals(2, Rule.parsePhonemeExpr("(|())").size());
+        assertEquals(3, Rule.parsePhonemeExpr("(|()|)").size());
+    }
+
     @Test
     void testSubSequenceWorks() {
         // AppendableCharSequence is private to Rule. We can only make it through a Phoneme.