From 9846ee3b54f351dfa4667860ebbf0ea68cd47cee Mon Sep 17 00:00:00 2001
From: Andy Heninger <andy.heninger@gmail.com>
Date: Wed, 18 Jan 2017 19:42:33 +0000
Subject: [PATCH] ICU-12918 Dictionary Break Iterator Assertion Failure

X-SVN-Rev: 39573
---
 icu4c/source/common/dictbe.cpp         | 19 ++++++++++++++++---
 icu4c/source/test/intltest/rbbitst.cpp | 21 +++++++++++++++++++++
 icu4c/source/test/intltest/rbbitst.h   |  1 +
 3 files changed, 38 insertions(+), 3 deletions(-)

diff --git a/icu4c/source/common/dictbe.cpp b/icu4c/source/common/dictbe.cpp
index 924f09bd5dd..93cb57cd4dc 100644
--- a/icu4c/source/common/dictbe.cpp
+++ b/icu4c/source/common/dictbe.cpp
@@ -1385,12 +1385,25 @@ CjkBreakEngine::divideUpDictionaryRange( UText *inText,
     // Now that we're done, convert positions in t_boundary[] (indices in 
     // the normalized input string) back to indices in the original input UText
     // while reversing t_boundary and pushing values to foundBreaks.
+    int32_t prevCPPos = -1;
+    int32_t prevUTextPos = -1;
     for (int32_t i = numBreaks-1; i >= 0; i--) {
         int32_t cpPos = t_boundary.elementAti(i);
+        U_ASSERT(cpPos > prevCPPos);
         int32_t utextPos =  inputMap.isValid() ? inputMap->elementAti(cpPos) : cpPos + rangeStart;
-        // Boundaries are added to foundBreaks output in ascending order.
-        U_ASSERT(foundBreaks.size() == 0 ||foundBreaks.peeki() < utextPos);
-        foundBreaks.push(utextPos, status);
+        U_ASSERT(utextPos >= prevUTextPos);
+        if (utextPos > prevUTextPos) {
+            // Boundaries are added to foundBreaks output in ascending order.
+            U_ASSERT(foundBreaks.size() == 0 || foundBreaks.peeki() < utextPos);
+            foundBreaks.push(utextPos, status);
+        } else {
+            // Normalization expanded the input text, the dictionary found a boundary
+            // within the expansion, giving two boundaries with the same index in the
+            // original text. Ignore the second. See ticket #12918.
+            --numBreaks;
+        }
+        prevCPPos = cpPos;
+        prevUTextPos = utextPos;
     }
 
     // inString goes out of scope
diff --git a/icu4c/source/test/intltest/rbbitst.cpp b/icu4c/source/test/intltest/rbbitst.cpp
index a3102c12d84..96a2b7a9edd 100644
--- a/icu4c/source/test/intltest/rbbitst.cpp
+++ b/icu4c/source/test/intltest/rbbitst.cpp
@@ -103,6 +103,7 @@ void RBBITest::runIndexedTest( int32_t index, UBool exec, const char* &name, cha
     TESTCASE_AUTO(TestBug5532);
     TESTCASE_AUTO(TestBug7547);
     TESTCASE_AUTO(TestBug12797);
+    TESTCASE_AUTO(TestBug12918);
     TESTCASE_AUTO_END;
 }
 
@@ -4652,6 +4653,26 @@ void RBBITest::TestBug12797() {
     }
 }
 
+void RBBITest::TestBug12918() {
+    // This test triggers an assertion failure in dictbe.cpp
+    const UChar *crasherString = u"\u3325\u4a16";
+    UErrorCode status = U_ZERO_ERROR;
+    UBreakIterator* iter = ubrk_open(UBRK_WORD, NULL, crasherString, -1, &status);
+    if (U_FAILURE(status)) {
+        errln("%s:%d status = %s", __FILE__, __LINE__, u_errorName(status));
+        return;
+    }
+    ubrk_first(iter);
+    int32_t pos = 0;
+    int32_t lastPos = -1;
+    while((pos = ubrk_next(iter)) != UBRK_DONE) {
+        if (pos <= lastPos) {
+            errln("%s:%d (pos, lastPos) = (%d, %d)", __FILE__, __LINE__, pos, lastPos);
+            break;
+        }
+    }
+    ubrk_close(iter);
+}
 
 //
 //  TestDebug    -  A place-holder test for debugging purposes.
diff --git a/icu4c/source/test/intltest/rbbitst.h b/icu4c/source/test/intltest/rbbitst.h
index 6b2c2f0eb72..91c3d99f633 100644
--- a/icu4c/source/test/intltest/rbbitst.h
+++ b/icu4c/source/test/intltest/rbbitst.h
@@ -76,6 +76,7 @@ public:
     void TestBug9983();
     void TestBug7547();
     void TestBug12797();
+    void TestBug12918();
 
     void TestDebug();
     void TestProperties();