From 08eae25b649efd039633d0a3f86b3a137d92b05a Mon Sep 17 00:00:00 2001
From: thadh <thadh@THADH-MTV.ad.corp.google.com>
Date: Sat, 9 May 2009 09:35:00 -0700
Subject: [PATCH] go

---
 .../android/dictionary/DictionaryBuilder.java |  7 ++-
 .../android/dictionary/DictionaryTest.java    | 50 +++++++++++++++++--
 2 files changed, 52 insertions(+), 5 deletions(-)

diff --git a/src/com/hughes/android/dictionary/DictionaryBuilder.java b/src/com/hughes/android/dictionary/DictionaryBuilder.java
index c141eff..820298e 100755
--- a/src/com/hughes/android/dictionary/DictionaryBuilder.java
+++ b/src/com/hughes/android/dictionary/DictionaryBuilder.java
@@ -28,9 +28,14 @@ public class DictionaryBuilder {
     }
     final String dictOutFilename = args[0];
 
-    final Dictionary dict = new Dictionary(Language.DE, Language.EN);
+    final Dictionary dict = new Dictionary("de-en.txt - a German-English dictionary\n" +
+    		"Version: 1.6, 2009-04-16\n" +
+    		"Source: http://dict.tu-chemnitz.de/\n" +
+    		"Thanks to Frank Richter.", Language.DE, Language.EN);
     System.out.println(Charset.forName("Cp1252"));
     processInputFile("c:\\de-en-chemnitz.txt", dict, true, Charset.forName("UTF8"));
+    
+    // Thad's extra sauce: 
     processInputFile("c:\\de-en-dictcc.txt", dict, false, Charset.forName("Cp1252"));
 
     createIndex(dict, Entry.LANG1);
diff --git a/src/com/hughes/android/dictionary/DictionaryTest.java b/src/com/hughes/android/dictionary/DictionaryTest.java
index 60a9cba..c09338a 100755
--- a/src/com/hughes/android/dictionary/DictionaryTest.java
+++ b/src/com/hughes/android/dictionary/DictionaryTest.java
@@ -3,8 +3,11 @@ package com.hughes.android.dictionary;
 import java.io.File;
 import java.io.IOException;
 import java.io.RandomAccessFile;
+import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.Collections;
 import java.util.List;
+import java.util.Random;
 import java.util.concurrent.atomic.AtomicBoolean;
 
 import junit.framework.TestCase;
@@ -29,7 +32,7 @@ public class DictionaryTest extends TestCase {
         Entry.parseFromLine("rennen :: run", false));
 
     {
-      final Dictionary dict = new Dictionary(Language.DE, Language.EN);
+      final Dictionary dict = new Dictionary("test", Language.DE, Language.EN);
       dict.entries.addAll(entries);
       DictionaryBuilder.createIndex(dict, Entry.LANG1);
       DictionaryBuilder.createIndex(dict, Entry.LANG2);
@@ -46,7 +49,7 @@ public class DictionaryTest extends TestCase {
     assertEquals(entries, dict.entries);
     
     assertEquals("der", dict.languageDatas[0].sortedIndex.get(0).word);
-    assertEquals("Die", dict.languageDatas[0].sortedIndex.get(1).word);
+    assertEquals("die", dict.languageDatas[0].sortedIndex.get(1).word);
     
     for (final IndexEntry indexEntry : dict.languageDatas[0].sortedIndex) {
       System.out.println(indexEntry);
@@ -75,7 +78,8 @@ public class DictionaryTest extends TestCase {
       }
     }
     
-    assertEquals("Die", dict.languageDatas[0].sortedIndex.get(dict.languageDatas[0].lookup("die", new AtomicBoolean())).word);
+    assertEquals("die", dict.languageDatas[0].sortedIndex.get(dict.languageDatas[0].lookup("Die", new AtomicBoolean())).word);
+    assertEquals("die", dict.languageDatas[0].sortedIndex.get(dict.languageDatas[0].lookup("die", new AtomicBoolean())).word);
 
   }
   
@@ -100,7 +104,7 @@ public class DictionaryTest extends TestCase {
 
     // Hyphenated words get put both multiple listings.
 
-    final Dictionary dict = new Dictionary(Language.DE, Language.EN);
+    final Dictionary dict = new Dictionary("test", Language.DE, Language.EN);
     dict.entries.addAll(entries);
     DictionaryBuilder.createIndex(dict, Entry.LANG1);
     DictionaryBuilder.createIndex(dict, Entry.LANG2);
@@ -115,6 +119,44 @@ public class DictionaryTest extends TestCase {
     }
 
   }
+  
+  public void testGermanSort() {
+    assertEquals("grosformat", Language.DE.normalizeTokenForSort("Grosformat"));
+    final List<String> words = Arrays.asList(
+        "er-ben",
+        "erben",
+        "Erben",
+        "Erbse",
+        "Erbsen",
+        "essen",
+        "Essen",
+        "Grosformat",
+        "Grosformats",
+        "Grossformat",
+        "Großformat",
+        "Grossformats",
+        "Großformats",
+        "Großpoo",
+        "Großpoos",
+        "hulle",
+        "Hulle",
+        "Hum",
+        "huelle",
+        "Huelle",
+        "hülle",
+        "Hülle"
+        );
+    for (final String s : words) {
+      System.out.println(s + "\t" + Language.DE.normalizeTokenForSort(s));
+    }
+    final List<String> shuffled = new ArrayList<String>(words);
+    Collections.shuffle(shuffled, new Random(0));
+    Collections.sort(shuffled, Language.DE.tokenComparator);
+    System.out.println(shuffled.toString());
+    for (int i = 0; i < words.size(); ++i) {
+      assertEquals(words.get(i), shuffled.get(i));
+    }
+  }
 
 
 }
-- 
2.43.0