src/third_party/icu/patches/utext.patch

diff --git a/source/common/utext.cpp b/source/common/utext.cpp
index 6dc31a0..0c7dd1e 100644
--- a/source/common/utext.cpp
+++ b/source/common/utext.cpp
@@ -1,7 +1,7 @@
 /*
 *******************************************************************************
 *
-*   Copyright (C) 2005-2015, International Business Machines
+*   Copyright (C) 2005-2016, International Business Machines
 *   Corporation and others.  All Rights Reserved.
 *
 *******************************************************************************
@@ -2524,6 +2524,7 @@ ucstrTextExtract(UText *ut,
             ut->chunkLength         = si;
             ut->nativeIndexingLimit = si;
             strLength               = si;
+            limit32                 = si;
             break;
         }
         U_ASSERT(di>=0); /* to ensure di never exceeds INT32_MAX, which must not happen logically */
@@ -2545,16 +2546,21 @@ ucstrTextExtract(UText *ut,
     // If the limit index points to a lead surrogate of a pair,
     //   add the corresponding trail surrogate to the destination.
     if (si>0 && U16_IS_LEAD(s[si-1]) &&
-        ((si<strLength || strLength<0)  && U16_IS_TRAIL(s[si])))
+            ((si<strLength || strLength<0)  && U16_IS_TRAIL(s[si])))
     {
         if (di<destCapacity) {
             // store only if there is space in the output buffer.
-            dest[di++] = s[si++];
+            dest[di++] = s[si];
         }
+        si++;
     }
 
     // Put iteration position at the point just following the extracted text
-    ut->chunkOffset = uprv_min(strLength, start32 + destCapacity);
+    if (si <= ut->chunkNativeLimit) {
+        ut->chunkOffset = si;
+    } else {
+        ucstrTextAccess(ut, si, TRUE);
+    }
 
     // Add a terminating NUL if space in the buffer permits,
     // and set the error status as required.
diff --git a/source/test/intltest/utxttest.cpp b/source/test/intltest/utxttest.cpp
index d8bedb6..801337c 100644
--- a/source/test/intltest/utxttest.cpp
+++ b/source/test/intltest/utxttest.cpp
@@ -1,6 +1,6 @@
 /********************************************************************
  * COPYRIGHT:
- * Copyright (c) 2005-2014, International Business Machines Corporation and
+ * Copyright (c) 2005-2016, International Business Machines Corporation and
  * others. All Rights Reserved.
  ********************************************************************/
 /************************************************************************
@@ -16,6 +16,8 @@
 #include "unicode/utf8.h"
 #include "unicode/ustring.h"
 #include "unicode/uchriter.h"
+#include "cmemory.h"
+#include "cstr.h"
 #include "utxttest.h"
 
 static UBool  gFailed = FALSE;
@@ -61,6 +63,8 @@ UTextTest::runIndexedTest(int32_t index, UBool exec,
             if (exec) Ticket10562();  break;
         case 6: name = "Ticket10983";
             if (exec) Ticket10983();  break;
+        case 7: name = "Ticket12130";
+            if (exec) Ticket12130(); break;
         default: name = "";          break;
     }
 }
@@ -1501,3 +1505,79 @@ void UTextTest::Ticket10983() {
 
     utext_close(ut);
 }
+
+// Ticket 12130 - extract on a UText wrapping a null terminated UChar * string
+//                leaves the iteration position set incorrectly when the
+//                actual string length is not yet known.
+//
+//                The test text needs to be long enough that UText defers getting the length.
+
+void UTextTest::Ticket12130() {
+    UErrorCode status = U_ZERO_ERROR;
+    
+    const char *text8 =
+        "Fundamentally, computers just deal with numbers. They store letters and other characters "
+        "by assigning a number for each one. Before Unicode was invented, there were hundreds "
+        "of different encoding systems for assigning these numbers. No single encoding could "
+        "contain enough characters: for example, the European Union alone requires several "
+        "different encodings to cover all its languages. Even for a single language like "
+        "English no single encoding was adequate for all the letters, punctuation, and technical "
+        "symbols in common use.";
+
+    UnicodeString str(text8);
+    const UChar *ustr = str.getTerminatedBuffer();
+    UText ut = UTEXT_INITIALIZER;
+    utext_openUChars(&ut, ustr, -1, &status);
+    UChar extractBuffer[50];
+
+    for (int32_t startIdx = 0; startIdx<str.length(); ++startIdx) {
+        int32_t endIdx = startIdx + 20;
+
+        u_memset(extractBuffer, 0, UPRV_LENGTHOF(extractBuffer));
+        utext_extract(&ut, startIdx, endIdx, extractBuffer, UPRV_LENGTHOF(extractBuffer), &status);
+        if (U_FAILURE(status)) {
+            errln("%s:%d %s", __FILE__, __LINE__, u_errorName(status));
+            return;
+        }
+        int64_t ni  = utext_getNativeIndex(&ut);
+        int64_t expectedni = startIdx + 20;
+        if (expectedni > str.length()) {
+            expectedni = str.length();
+        }
+        if (expectedni != ni) {
+            errln("%s:%d utext_getNativeIndex() expected %d, got %d", __FILE__, __LINE__, expectedni, ni);
+        }
+        if (0 != str.tempSubString(startIdx, 20).compare(extractBuffer)) { 
+            errln("%s:%d utext_extract() failed. expected \"%s\", got \"%s\"",
+                    __FILE__, __LINE__, CStr(str.tempSubString(startIdx, 20))(), CStr(UnicodeString(extractBuffer))());
+        }
+    }
+    utext_close(&ut);
+
+    // Similar utext extract, this time with the string length provided to the UText in advance,
+    // and a buffer of larger than required capacity.
+   
+    utext_openUChars(&ut, ustr, str.length(), &status);
+    for (int32_t startIdx = 0; startIdx<str.length(); ++startIdx) {
+        int32_t endIdx = startIdx + 20;
+        u_memset(extractBuffer, 0, UPRV_LENGTHOF(extractBuffer));
+        utext_extract(&ut, startIdx, endIdx, extractBuffer, UPRV_LENGTHOF(extractBuffer), &status);
+        if (U_FAILURE(status)) {
+            errln("%s:%d %s", __FILE__, __LINE__, u_errorName(status));
+            return;
+        }
+        int64_t ni  = utext_getNativeIndex(&ut);
+        int64_t expectedni = startIdx + 20;
+        if (expectedni > str.length()) {
+            expectedni = str.length();
+        }
+        if (expectedni != ni) {
+            errln("%s:%d utext_getNativeIndex() expected %d, got %d", __FILE__, __LINE__, expectedni, ni);
+        }
+        if (0 != str.tempSubString(startIdx, 20).compare(extractBuffer)) { 
+            errln("%s:%d utext_extract() failed. expected \"%s\", got \"%s\"",
+                    __FILE__, __LINE__, CStr(str.tempSubString(startIdx, 20))(), CStr(UnicodeString(extractBuffer))());
+        }
+    }
+    utext_close(&ut);
+}
diff --git a/source/test/intltest/utxttest.h b/source/test/intltest/utxttest.h
index 7091655..346563d 100644
--- a/source/test/intltest/utxttest.h
+++ b/source/test/intltest/utxttest.h
@@ -1,6 +1,6 @@
 /********************************************************************
  * COPYRIGHT: 
- * Copyright (c) 2005-2014, International Business Machines Corporation and
+ * Copyright (c) 2005-2016, International Business Machines Corporation and
  * others. All Rights Reserved.
  ********************************************************************/
 /************************************************************************
@@ -35,6 +35,7 @@ public:
     void Ticket6847();
     void Ticket10562();
     void Ticket10983();
+    void Ticket12130();
 
 private:
     struct m {                              // Map between native indices & code points.