LUCENE-8596: Treat hash mark as comment only at beginning of line in kuromoji

user dictionary. Via Masaru Hasegawa and Satoshi Kato

lucene/CHANGES.txt

@@ -58,6 +58,10 @@ Improvements
   
 * LUCENE-8984: MoreLikeThis MLT is biased for uncommon fields (Andy Hind via Anshum Gupta)
 
+* LUCENE-8596: Kuromoji user dictionary now accepts entries containing hash mark (#) that were
+  previously treated as beginning a line-ending comment (Satoshi Kato and Masaru Hasegawa via
+  Michael Sokolov)
+
 Bug fixes
 
 * LUCENE-8663: NRTCachingDirectory.slowFileExists may open a file while 

lucene/analysis/kuromoji/src/java/org/apache/lucene/analysis/ja/dict/UserDictionary.java

@@ -65,7 +65,7 @@ public final class UserDictionary implements Dictionary {
     // text, segmentation, readings, POS
     while ((line = br.readLine()) != null) {
       // Remove comments
-      line = line.replaceAll("#.*$", "");
+      line = line.replaceAll("^#.*$", "");
 
       // Skip empty lines or comment lines
       if (line.trim().length() == 0) {

lucene/analysis/kuromoji/src/test/org/apache/lucene/analysis/ja/dict/UserDictionaryTest.java

@@ -99,4 +99,16 @@ public class UserDictionaryTest extends LuceneTestCase {
     assertTrue(e.getMessage().contains("does not match the surface form"));
   }
 
+  @Test
+  public void testSharp() throws IOException {
+    String[] inputs = {"テスト#", "テスト#テスト"};
+    UserDictionary dictionary = TestJapaneseTokenizer.readDict();
+
+    for (String input: inputs) {
+      System.out.println(input);
+      int[][] result = dictionary.lookup(input.toCharArray(), 0, input.length());
+      assertEquals("カスタム名刺", dictionary.getPartOfSpeech(result[0][0]));
+    }
+  }
+
 }

lucene/analysis/kuromoji/src/test/org/apache/lucene/analysis/ja/userdict.txt

@@ -8,3 +8,7 @@
 # Silly entry:
 abcd,a b cd,foo1 foo2 foo3,bar
 abcdefg,ab cd efg,foo1 foo2 foo4,bar
+
+# sharp test
+test#テスト,test # テスト,test # テスト,カスタム名刺
+テスト#,テスト #,テスト #,カスタム名刺