CJK segmentation patch for ICU 4.6...

icu46/source/test/intltest/rbbiapts.cpp

Index: icu46/source/test/intltest/rbbiapts.cpp
===================================================================
--- icu46/source/test/intltest/rbbiapts.cpp	(revision 68397)
+++ icu46/source/test/intltest/rbbiapts.cpp	(working copy)
@@ -156,9 +156,13 @@
     if(*a!=*b){
         errln("Failed: boilerplate method operator!= does not return correct results");
     }
-    BreakIterator* c = BreakIterator::createWordInstance(Locale("ja"),status);
-    if(a && c){
-        if(*c==*a){
+    // Japanese word break iteratos is identical to root with
+    // a dictionary-based break iterator, but Thai character break iterator
+    // is still different from Root. 
+    BreakIterator* c = BreakIterator::createCharacterInstance(Locale("ja"),status);
+    BreakIterator* d = BreakIterator::createCharacterInstance(Locale("th"),status);
+    if(c && d){
+        if(*c==*d){
             errln("Failed: boilerplate method opertator== does not return correct results");
         }
     }else{
@@ -167,6 +171,7 @@
     delete a;
     delete b;
     delete c;
+    delete d;
 }
 
 void RBBIAPITest::TestgetRules()
@@ -635,21 +640,21 @@
 //
 void RBBIAPITest::TestRuleStatus() {
      UChar str[30];
-     u_unescape("plain word 123.45 \\u9160\\u9161 \\u30a1\\u30a2 \\u3041\\u3094",
-              // 012345678901234567  8      9    0  1      2    3  4      5    6
-              //                    Ideographic    Katakana       Hiragana
+     //no longer test Han or hiragana breaking here: ruleStatusVec would return nothing
+     // changed UBRK_WORD_KANA to UBRK_WORD_IDEO
+     u_unescape("plain word 123.45 \\u30a1\\u30a2 ",
+              // 012345678901234567  8      9    0      
+              //                     Katakana      
                 str, 30);
      UnicodeString testString1(str);
-     int32_t bounds1[] = {0, 5, 6, 10, 11, 17, 18, 19, 20, 21, 23, 24, 25, 26};
+     int32_t bounds1[] = {0, 5, 6, 10, 11, 17, 18, 20, 21};
      int32_t tag_lo[]  = {UBRK_WORD_NONE,     UBRK_WORD_LETTER, UBRK_WORD_NONE,    UBRK_WORD_LETTER,
                           UBRK_WORD_NONE,     UBRK_WORD_NUMBER, UBRK_WORD_NONE,
-                          UBRK_WORD_IDEO,     UBRK_WORD_IDEO,   UBRK_WORD_NONE,
-                          UBRK_WORD_KANA,     UBRK_WORD_NONE,   UBRK_WORD_KANA,    UBRK_WORD_KANA};
+                          UBRK_WORD_IDEO,     UBRK_WORD_NONE};
 
      int32_t tag_hi[]  = {UBRK_WORD_NONE_LIMIT, UBRK_WORD_LETTER_LIMIT, UBRK_WORD_NONE_LIMIT, UBRK_WORD_LETTER_LIMIT,
                           UBRK_WORD_NONE_LIMIT, UBRK_WORD_NUMBER_LIMIT, UBRK_WORD_NONE_LIMIT,
-                          UBRK_WORD_IDEO_LIMIT, UBRK_WORD_IDEO_LIMIT,   UBRK_WORD_NONE_LIMIT,
-                          UBRK_WORD_KANA_LIMIT, UBRK_WORD_NONE_LIMIT,   UBRK_WORD_KANA_LIMIT, UBRK_WORD_KANA_LIMIT};
+                          UBRK_WORD_IDEO_LIMIT, UBRK_WORD_NONE_LIMIT};
 
      UErrorCode status=U_ZERO_ERROR;
 
@@ -888,9 +893,11 @@
 
     URegistryKey key = BreakIterator::registerInstance(ja_word, "xx", UBRK_WORD, status);
     {
+#if 0 // With a dictionary based word breaking, ja_word is identical to root.
         if (ja_word && *ja_word == *root_word) {
             errln("japan not different from root");
         }
+#endif
     }
 
     {