From 4d90dc692d684cba36bc30404d80f6f6c6dfbe25 Mon Sep 17 00:00:00 2001
From: Andy Heninger <andy.heninger@gmail.com>
Date: Thu, 10 Sep 2009 23:17:38 +0000
Subject: [PATCH] ICU-7130 infinite loop in dictionary break iterators

X-SVN-Rev: 26618
---
 icu4c/source/common/rbbi.cpp           |  9 ++++---
 icu4c/source/test/intltest/rbbitst.cpp | 37 ++++++++++++++++++++++++++
 icu4c/source/test/intltest/rbbitst.h   |  1 +
 3 files changed, 44 insertions(+), 3 deletions(-)

diff --git a/icu4c/source/common/rbbi.cpp b/icu4c/source/common/rbbi.cpp
index b06a01c093..5308e26d09 100644
--- a/icu4c/source/common/rbbi.cpp
+++ b/icu4c/source/common/rbbi.cpp
@@ -1,6 +1,6 @@
 /*
 ***************************************************************************
-*   Copyright (C) 1999-2008 International Business Machines Corporation   *
+*   Copyright (C) 1999-2009 International Business Machines Corporation   *
 *   and others. All rights reserved.                                      *
 ***************************************************************************
 */
@@ -247,7 +247,10 @@ void RuleBasedBreakIterator::init() {
     fLastRuleStatusIndex  = 0;
     fLastStatusIndexValid = TRUE;
     fDictionaryCharCount  = 0;
-    fBreakType            = -1;
+    fBreakType            = UBRK_WORD;  // Defaulting BreakType to word gives reasonable
+                                        //   dictionary behavior for Break Iterators that are
+                                        //   built from rules.  Even better would be the ability to
+                                        //   declare the type in the rules.
 
     fCachedBreakPositions    = NULL;
     fLanguageBreakEngines    = NULL;
@@ -1019,7 +1022,7 @@ int32_t RuleBasedBreakIterator::handleNext(const RBBIStateTable *statetable) {
 
         #ifdef RBBI_DEBUG
             if (fTrace) {
-                RBBIDebugPrintf("             %4d   ", utext_getNativeIndex(fText));
+                RBBIDebugPrintf("             %4ld   ", utext_getNativeIndex(fText));
                 if (0x20<=c && c<0x7f) {
                     RBBIDebugPrintf("\"%c\"  ", c);
                 } else {
diff --git a/icu4c/source/test/intltest/rbbitst.cpp b/icu4c/source/test/intltest/rbbitst.cpp
index 4f0acd68ca..efbb7c1c51 100644
--- a/icu4c/source/test/intltest/rbbitst.cpp
+++ b/icu4c/source/test/intltest/rbbitst.cpp
@@ -140,6 +140,8 @@ void RBBITest::runIndexedTest( int32_t index, UBool exec, const char* &name, cha
         case 21: case 22: case 23: name = "skip";
             break;
 #endif
+        case 24: name = "TestDictRules";
+            if (exec) TestDictRules();                         break;
 
         default: name = ""; break; //needed to end loop
     }
@@ -1991,6 +1993,41 @@ void RBBITest::TestTailoredBreaks() {
 }
 
 
+//-------------------------------------------------------------------------------
+//
+//  TestDictRules   create a break iterator from source rules that includes a
+//                  dictionary range.   Regression for bug #7130.  Source rules
+//                  do not declare a break iterator type (word, line, sentence, etc.
+//                  but the dictionary code, without a type, would loop.
+//
+//-------------------------------------------------------------------------------
+void RBBITest::TestDictRules() {
+    const char *rules =  "$dictionary = [a-z]; \n"
+                         "!!forward; \n"
+                         "$dictionary $dictionary; \n"
+                         "!!reverse; \n"
+                         "$dictionary $dictionary; \n";
+    const char *text = "aa";
+    UErrorCode status = U_ZERO_ERROR;
+    UParseError parseError;
+
+    RuleBasedBreakIterator bi(rules, parseError, status);
+    TEST_ASSERT_SUCCESS(status);
+    UnicodeString utext = text;
+    bi.setText(utext);
+    int32_t position;
+    int32_t loops;
+    for (loops = 0; loops<10; loops++) {
+        position = bi.next();
+        if (position == RuleBasedBreakIterator::DONE) {
+            break;
+        }
+    }
+    TEST_ASSERT(loops == 1);
+}
+
+
+
 //-------------------------------------------------------------------------------
 //
 //    ReadAndConvertFile   Read a text data file, convert it to UChars, and
diff --git a/icu4c/source/test/intltest/rbbitst.h b/icu4c/source/test/intltest/rbbitst.h
index 69511bd85d..fbdbbf880e 100644
--- a/icu4c/source/test/intltest/rbbitst.h
+++ b/icu4c/source/test/intltest/rbbitst.h
@@ -70,6 +70,7 @@ public:
     void TestBug5775();
     void TestThaiBreaks();
     void TestTailoredBreaks();
+    void TestDictRules();
 
     void TestDebug();