From 20ae9261947dce1b705ee80eacad85c34d72e58b Mon Sep 17 00:00:00 2001
From: Peter Edberg <pedberg@unicode.org>
Date: Sun, 19 May 2013 06:36:13 +0000
Subject: [PATCH] ICU-10120 Copy fix #9983 (for word ubrk_previous infinite
 loop) to word_POSIX too

X-SVN-Rev: 33682
---
 icu4c/source/data/brkitr/word_POSIX.txt |  4 ++--
 icu4c/source/test/intltest/rbbitst.cpp  | 22 +++++++++++++++++++---
 icu4c/source/test/testdata/rbbitst.txt  |  1 +
 3 files changed, 22 insertions(+), 5 deletions(-)

diff --git a/icu4c/source/data/brkitr/word_POSIX.txt b/icu4c/source/data/brkitr/word_POSIX.txt
index 3cd0372db66..fbab7573e30 100644
--- a/icu4c/source/data/brkitr/word_POSIX.txt
+++ b/icu4c/source/data/brkitr/word_POSIX.txt
@@ -96,7 +96,7 @@ $CR $LF;
 #          begins with a group of Format chars, or with a "word" consisting of a single
 #          char that is not in any of the listed word break categories followed by
 #          format char(s), or is not a CJK dictionary character.
-[^$CR $LF $Newline $dictionaryCJK]? ($Extend |  $Format)+;
+[^$CR $LF $Newline]? ($Extend |  $Format)+;
 
 $NumericEx {100};
 $ALetterEx {200};
@@ -186,7 +186,7 @@ $BackRegional_IndicatorEx = ($Format | $Extend)* $Regional_Indicator;
 $LF $CR;
 
 # rule 4
-($Format | $Extend)*  [^$CR $LF $Newline $dictionaryCJK]?;
+($Format | $Extend)*  [^$CR $LF $Newline]?;
 
 # rule 5
 
diff --git a/icu4c/source/test/intltest/rbbitst.cpp b/icu4c/source/test/intltest/rbbitst.cpp
index c7f66aa8e21..00cc6d0f048 100644
--- a/icu4c/source/test/intltest/rbbitst.cpp
+++ b/icu4c/source/test/intltest/rbbitst.cpp
@@ -4261,18 +4261,34 @@ void RBBITest::TestBug9983(void)  {
     LocalPointer<RuleBasedBreakIterator> brkiter(static_cast<RuleBasedBreakIterator *>(
         BreakIterator::createWordInstance(Locale::getRoot(), status)));
     TEST_ASSERT_SUCCESS(status);
+    LocalPointer<RuleBasedBreakIterator> brkiterPOSIX(static_cast<RuleBasedBreakIterator *>(
+        BreakIterator::createWordInstance(Locale::createFromName("en_US_POSIX"), status)));
+    TEST_ASSERT_SUCCESS(status);
     if (U_FAILURE(status)) {
         return;
     }
+    int32_t offset, rstatus, iterationCount;
+
     brkiter->setText(text);
-    int32_t offset, rstatus;
     brkiter->last();
-    int32_t iterationCount = 0;
+    iterationCount = 0;
     while ( (offset = brkiter->previous()) != UBRK_DONE ) {
         iterationCount++;
         rstatus = brkiter->getRuleStatus();
         (void)rstatus;     // Suppress set but not used warning.
-        // printf(" %d(%d)", offset, rstatus);
+        if (iterationCount >= 10) {
+           break; 
+        }
+    }
+    TEST_ASSERT(iterationCount == 6);
+
+    brkiterPOSIX->setText(text);
+    brkiterPOSIX->last();
+    iterationCount = 0;
+    while ( (offset = brkiterPOSIX->previous()) != UBRK_DONE ) {
+        iterationCount++;
+        rstatus = brkiterPOSIX->getRuleStatus();
+        (void)rstatus;     // Suppress set but not used warning.
         if (iterationCount >= 10) {
            break; 
         }
diff --git a/icu4c/source/test/testdata/rbbitst.txt b/icu4c/source/test/testdata/rbbitst.txt
index 5dde3f8c951..4231d271f60 100644
--- a/icu4c/source/test/testdata/rbbitst.txt
+++ b/icu4c/source/test/testdata/rbbitst.txt
@@ -790,6 +790,7 @@ Bangkok)•</data>
 <word>
 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx<200>:•yy<200> •or<200> •struct<200>.•field<200> \
 •for<200> •CS<200>-•types<200>.•</data>
+<data>•\u06c9<200>\uc799\ufffa•</data>
 
 
 # UBreakIteratorType UBRK_CHARACTER, Locale "th"
-- 
2.49.0