From: Maxime Serrano <mserrano@svn.icu-project.org>
Date: Fri, 17 Aug 2012 21:27:08 +0000 (+0000)
Subject: ICU-9353 Do not generate surrogates in the Monkey test so as to avoid generating... 
X-Git-Tag: milestone-59-0-1~3652
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=1a702c375507868a5c28e97f77671fc653175f03;p=icu

ICU-9353 Do not generate surrogates in the Monkey test so as to avoid generating higher CJK characters; also, hand off to Unhandled if we do not have a breakengine of this type

X-SVN-Rev: 32195
---

diff --git a/icu4j/main/classes/core/src/com/ibm/icu/text/RuleBasedBreakIterator.java b/icu4j/main/classes/core/src/com/ibm/icu/text/RuleBasedBreakIterator.java
index 787f76ca7c1..4ff71987431 100644
--- a/icu4j/main/classes/core/src/com/ibm/icu/text/RuleBasedBreakIterator.java
+++ b/icu4j/main/classes/core/src/com/ibm/icu/text/RuleBasedBreakIterator.java
@@ -1036,12 +1036,21 @@ public class RuleBasedBreakIterator extends BreakIterator {
             case UScript.KATAKANA:
             case UScript.HIRAGANA:
             case UScript.HAN:
-                if (getBreakType() == KIND_WORD)
+                if (getBreakType() == KIND_WORD) {
                     eng = new CjkBreakEngine(false);
+                }
+                else {
+                    fUnhandledBreakEngine.handleChar(c, getBreakType());
+                    eng = fUnhandledBreakEngine;
+                }
                 break;
             case UScript.HANGUL:
-                if (getBreakType() == KIND_WORD)
+                if (getBreakType() == KIND_WORD) {
                     eng = new CjkBreakEngine(true);
+                } else {
+                    fUnhandledBreakEngine.handleChar(c, getBreakType());
+                    eng = fUnhandledBreakEngine;
+                }
                 break;
             default:
                 fUnhandledBreakEngine.handleChar(c, getBreakType());
diff --git a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/rbbi/RBBITestMonkey.java b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/rbbi/RBBITestMonkey.java
index 6b3c1308bd3..9e3719b5348 100644
--- a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/rbbi/RBBITestMonkey.java
+++ b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/rbbi/RBBITestMonkey.java
@@ -300,7 +300,8 @@ public class RBBITestMonkey extends TestFmwk {
             fOtherSet.removeAll(fExtendSet);
             fOtherSet.removeAll(fExtendNumLetSet);
             // Inhibit dictionary characters from being tested at all.
-            fOtherSet.removeAll(new UnicodeSet("[\\p{LineBreak = Complex_Context}]"));
+            // remove surrogates so as to not generate higher CJK characters
+            fOtherSet.removeAll(new UnicodeSet("[[\\p{LineBreak = Complex_Context}][:Line_Break=Surrogate:]]"));
             fOtherSet.removeAll(fDictionaryCjkSet);
 
             fSets            = new ArrayList();
@@ -613,7 +614,6 @@ public class RBBITestMonkey extends TestFmwk {
             fSets.add(fWJ);
             fSets.add(fSA);
             fSets.add(fSG);
-            
         }
         
         void setText(StringBuffer s) {