X-Git-Url: https://git.saurik.com/apple/icu.git/blobdiff_plain/73c04bcfe1096173b00431f0cdc742894b15eef0..2be6500137966bba13903b24204a44804f9e133a:/icuSources/test/intltest/ustrtest.cpp

diff --git a/icuSources/test/intltest/ustrtest.cpp b/icuSources/test/intltest/ustrtest.cpp
index 17c7b2b7..6a27ea3c 100644
--- a/icuSources/test/intltest/ustrtest.cpp
+++ b/icuSources/test/intltest/ustrtest.cpp
@@ -1,27 +1,27 @@
 /********************************************************************
  * COPYRIGHT: 
- * Copyright (c) 1997-2004, International Business Machines Corporation and
+ * Copyright (c) 1997-2012, International Business Machines Corporation and
  * others. All Rights Reserved.
  ********************************************************************/
 
 #include "ustrtest.h"
+#include "unicode/appendable.h"
+#include "unicode/std_string.h"
 #include "unicode/unistr.h"
 #include "unicode/uchar.h"
 #include "unicode/ustring.h"
 #include "unicode/locid.h"
 #include "unicode/ucnv.h"
+#include "unicode/uenum.h"
+#include "unicode/utf16.h"
 #include "cmemory.h"
 #include "charstr.h"
 
 #if 0
 #include "unicode/ustream.h"
 
-#if U_IOSTREAM_SOURCE >= 199711
 #include <iostream>
 using namespace std;
-#elif U_IOSTREAM_SOURCE >= 198506
-#include <iostream.h>
-#endif
 
 #endif
 
@@ -56,7 +56,14 @@ void UnicodeStringTest::runIndexedTest( int32_t index, UBool exec, const char* &
         case 13: name = "TestUnescape"; if (exec) TestUnescape(); break;
         case 14: name = "TestCountChar32"; if (exec) TestCountChar32(); break;
         case 15: name = "TestStringEnumeration"; if (exec) TestStringEnumeration(); break;
-        case 16: name = "TestCharString"; if (exec) TestCharString(); break;
+        case 16: name = "TestNameSpace"; if (exec) TestNameSpace(); break;
+        case 17: name = "TestUTF32"; if (exec) TestUTF32(); break;
+        case 18: name = "TestUTF8"; if (exec) TestUTF8(); break;
+        case 19: name = "TestReadOnlyAlias"; if (exec) TestReadOnlyAlias(); break;
+        case 20: name = "TestAppendable"; if (exec) TestAppendable(); break;
+        case 21: name = "TestUnicodeStringImplementsAppendable"; if (exec) TestUnicodeStringImplementsAppendable(); break;
+        case 22: name = "TestSizeofUnicodeString"; if (exec) TestSizeofUnicodeString(); break;
+        case 23: name = "TestStartsWithAndEndsWithNulTerminated"; if (exec) TestStartsWithAndEndsWithNulTerminated(); break;
 
         default: name = ""; break; //needed to end loop
     }
@@ -106,9 +113,9 @@ UnicodeStringTest::TestBasicManipulation()
         errln("operator+=() failed:  expected \"" + expectedValue + "\"\n,got \"" + test2 + "\"");
     
     if (test1.length() != 70)
-        errln("length() failed: expected 70, got " + test1.length());
+        errln(UnicodeString("length() failed: expected 70, got ") + test1.length());
     if (test2.length() != 30)
-        errln("length() failed: expected 30, got " + test2.length());
+        errln(UnicodeString("length() failed: expected 30, got ") + test2.length());
 
     UnicodeString test3;
     test3.append((UChar32)0x20402);
@@ -116,7 +123,7 @@ UnicodeStringTest::TestBasicManipulation()
         errln((UnicodeString)"append failed for UChar32, expected \"\\\\ud841\\\\udc02\", got " + prettify(test3));
     }
     if(test3.length() != 2){
-        errln("append or length failed for UChar32, expected 2, got " + test3.length());
+        errln(UnicodeString("append or length failed for UChar32, expected 2, got ") + test3.length());
     }
     test3.append((UChar32)0x0074);
     if(test3 != CharsToUnicodeString("\\uD841\\uDC02t")){
@@ -224,10 +231,33 @@ UnicodeStringTest::TestBasicManipulation()
         cnv=ucnv_open("ISO-8859-1", &errorCode);
         UnicodeString v(cs, -1, cnv, errorCode);
         ucnv_close(cnv);
-        if(v!=UnicodeString("a\\xe4\\x85").unescape()) {
+        if(v!=CharsToUnicodeString("a\\xe4\\x85")) {
             errln("UnicodeString(const char *, length, cnv, errorCode) does not work with length==-1");
         }
     }
+
+#if U_CHARSET_IS_UTF8
+    {
+        // Test the hardcoded-UTF-8 UnicodeString optimizations.
+        static const uint8_t utf8[]={ 0x61, 0xC3, 0xA4, 0xC3, 0x9F, 0xE4, 0xB8, 0x80, 0 };
+        static const UChar utf16[]={ 0x61, 0xE4, 0xDF, 0x4E00 };
+        UnicodeString from8a = UnicodeString((const char *)utf8);
+        UnicodeString from8b = UnicodeString((const char *)utf8, (int32_t)sizeof(utf8)-1);
+        UnicodeString from16(FALSE, utf16, LENGTHOF(utf16));
+        if(from8a != from16 || from8b != from16) {
+            errln("UnicodeString(const char * U_CHARSET_IS_UTF8) failed");
+        }
+        char buffer[16];
+        int32_t length8=from16.extract(0, 0x7fffffff, buffer, (uint32_t)sizeof(buffer));
+        if(length8!=((int32_t)sizeof(utf8)-1) || 0!=uprv_memcmp(buffer, utf8, sizeof(utf8))) {
+            errln("UnicodeString::extract(char * U_CHARSET_IS_UTF8) failed");
+        }
+        length8=from16.extract(1, 2, buffer, (uint32_t)sizeof(buffer));
+        if(length8!=4 || buffer[length8]!=0 || 0!=uprv_memcmp(buffer, utf8+1, length8)) {
+            errln("UnicodeString::extract(substring to char * U_CHARSET_IS_UTF8) failed");
+        }
+    }
+#endif
 }
 
 void
@@ -691,14 +721,15 @@ UnicodeStringTest::TestSearching()
           (startPos = test1.indexOf(test2, startPos)) != -1 ? (++occurrences, startPos += 4) : 0)
         ;
     if (occurrences != 6)
-        errln("indexOf failed: expected to find 6 occurrences, found " + occurrences);
-    
+        errln(UnicodeString("indexOf failed: expected to find 6 occurrences, found ") + occurrences);
+
     for ( occurrences = 0, startPos = 10;
           startPos != -1 && startPos < test1.length();
           (startPos = test1.indexOf(test2, startPos)) != -1 ? (++occurrences, startPos += 4) : 0)
         ;
     if (occurrences != 4)
-        errln("indexOf with starting offset failed: expected to find 4 occurrences, found " + occurrences);
+        errln(UnicodeString("indexOf with starting offset failed: "
+                            "expected to find 4 occurrences, found ") + occurrences);
 
     int32_t endPos = 28;
     for ( occurrences = 0, startPos = 5;
@@ -706,7 +737,8 @@ UnicodeStringTest::TestSearching()
           (startPos = test1.indexOf(test2, startPos, endPos - startPos)) != -1 ? (++occurrences, startPos += 4) : 0)
         ;
     if (occurrences != 4)
-        errln("indexOf with starting and ending offsets failed: expected to find 4 occurrences, found " + occurrences);
+        errln(UnicodeString("indexOf with starting and ending offsets failed: "
+                            "expected to find 4 occurrences, found ") + occurrences);
 
     //using UChar32 string
     for ( startPos=0, occurrences=0;
@@ -721,7 +753,7 @@ UnicodeStringTest::TestSearching()
           (startPos = test3.indexOf(test4, startPos)) != -1 ? (++occurrences, startPos += 2) : 0)
         ;
     if (occurrences != 2)
-        errln("indexOf failed: expected to find 2 occurrences, found " + occurrences);
+        errln(UnicodeString("indexOf failed: expected to find 2 occurrences, found ") + occurrences);
     //---
 
     for ( occurrences = 0, startPos = 0;
@@ -729,21 +761,24 @@ UnicodeStringTest::TestSearching()
           (startPos = test1.indexOf(testChar, startPos)) != -1 ? (++occurrences, startPos += 1) : 0)
         ;
     if (occurrences != 16)
-        errln("indexOf with character failed: expected to find 16 occurrences, found " + occurrences);
+        errln(UnicodeString("indexOf with character failed: "
+                            "expected to find 16 occurrences, found ") + occurrences);
 
     for ( occurrences = 0, startPos = 10;
           startPos != -1 && startPos < test1.length();
           (startPos = test1.indexOf(testChar, startPos)) != -1 ? (++occurrences, startPos += 1) : 0)
         ;
     if (occurrences != 12)
-        errln("indexOf with character & start offset failed: expected to find 12 occurrences, found " + occurrences);
+        errln(UnicodeString("indexOf with character & start offset failed: "
+                            "expected to find 12 occurrences, found ") + occurrences);
 
     for ( occurrences = 0, startPos = 5, endPos = 28;
           startPos != -1 && startPos < test1.length();
           (startPos = test1.indexOf(testChar, startPos, endPos - startPos)) != -1 ? (++occurrences, startPos += 1) : 0)
         ;
     if (occurrences != 10)
-        errln("indexOf with character & start & end offsets failed: expected to find 10 occurrences, found " + occurrences);
+        errln(UnicodeString("indexOf with character & start & end offsets failed: "
+                            "expected to find 10 occurrences, found ") + occurrences);
 
     //testing for UChar32
     UnicodeString subString;
@@ -786,14 +821,16 @@ UnicodeStringTest::TestSearching()
           (startPos = test1.lastIndexOf(test2, 5, startPos - 5)) != -1 ? ++occurrences : 0)
         ;
     if (occurrences != 4)
-        errln("lastIndexOf with starting and ending offsets failed: expected to find 4 occurrences, found " + occurrences);
+        errln(UnicodeString("lastIndexOf with starting and ending offsets failed: "
+                            "expected to find 4 occurrences, found ") + occurrences);
 
     for ( occurrences = 0, startPos = 32;
           startPos != -1;
           (startPos = test1.lastIndexOf(testChar, 5, startPos - 5)) != -1 ? ++occurrences : 0)
         ;
     if (occurrences != 11)
-        errln("lastIndexOf with character & start & end offsets failed: expected to find 11 occurrences, found " + occurrences);
+        errln(UnicodeString("lastIndexOf with character & start & end offsets failed: "
+                            "expected to find 11 occurrences, found ") + occurrences);
 
     //testing UChar32
     startPos=test3.length();
@@ -915,27 +952,38 @@ UnicodeStringTest::TestPrefixAndSuffix()
         errln("endsWith() failed: \"" + test2 + "\" shouldn't be a suffix of \"" + test1 + "\".");
     }
 
-   if (!test1.endsWith(test3)) { 
-     errln("endsWith(test3) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
-   }
-   if (!test1.endsWith(test3, 0, INT32_MAX)) {
-     errln("endsWith(test3, 0, INT32_MAX) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
-   }
-
-   if(!test1.endsWith(test3.getBuffer(), test3.length())) {
-     errln("endsWith(test3.getBuffer(), test3.length()) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
-   }
-   if(!test1.endsWith(test3.getTerminatedBuffer(), 0, -1)) {
-     errln("endsWith(test3.getTerminatedBuffer(), 0, -1) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
-   }
-   
-   if (!test3.startsWith(test4)) {
-     errln("endsWith(test4) failed: \"" + test4 + "\" should be a prefix of \"" + test3 + "\".");
-   }
-   
-   if (test4.startsWith(test3)) {
-     errln("startsWith(test3) failed: \"" + test3 + "\" shouldn't be a prefix of \"" + test4 + "\".");
-   }
+    if (!test1.endsWith(test3)) { 
+        errln("endsWith(test3) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
+    }
+    if (!test1.endsWith(test3, 0, INT32_MAX)) {
+        errln("endsWith(test3, 0, INT32_MAX) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
+    }
+
+    if(!test1.endsWith(test3.getBuffer(), test3.length())) {
+        errln("endsWith(test3.getBuffer(), test3.length()) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
+    }
+    if(!test1.endsWith(test3.getTerminatedBuffer(), 0, -1)) {
+        errln("endsWith(test3.getTerminatedBuffer(), 0, -1) failed: \"" + test3 + "\" should be a suffix of \"" + test1 + "\".");
+    }
+
+    if (!test3.startsWith(test4)) {
+        errln("endsWith(test4) failed: \"" + test4 + "\" should be a prefix of \"" + test3 + "\".");
+    }
+
+    if (test4.startsWith(test3)) {
+        errln("startsWith(test3) failed: \"" + test3 + "\" shouldn't be a prefix of \"" + test4 + "\".");
+    }
+}
+
+void
+UnicodeStringTest::TestStartsWithAndEndsWithNulTerminated() {
+    UnicodeString test("abcde");
+    const UChar ab[] = { 0x61, 0x62, 0 };
+    const UChar de[] = { 0x64, 0x65, 0 };
+    assertTrue("abcde.startsWith(ab, -1)", test.startsWith(ab, -1));
+    assertTrue("abcde.startsWith(ab, 0, -1)", test.startsWith(ab, 0, -1));
+    assertTrue("abcde.endsWith(de, -1)", test.endsWith(de, -1));
+    assertTrue("abcde.endsWith(de, 0, -1)", test.endsWith(de, 0, -1));
 }
 
 void
@@ -978,6 +1026,16 @@ UnicodeStringTest::TestReverse()
     if(test.char32At(0)!=0x1ed0 || test.char32At(1)!=0xc4 || test.char32At(2)!=0x1d15f || test.char32At(4)!=0x2f999) {
         errln("reverse() failed with supplementary characters");
     }
+
+    // Test case for ticket #8091:
+    // UnicodeString::reverse() failed to see a lead surrogate in the middle of
+    // an odd-length string that contains no other lead surrogates.
+    test=UNICODE_STRING_SIMPLE("ab\\U0001F4A9e").unescape();
+    UnicodeString expected=UNICODE_STRING_SIMPLE("e\\U0001F4A9ba").unescape();
+    test.reverse();
+    if(test!=expected) {
+        errln("reverse() failed with only lead surrogate in the middle");
+    }
 }
 
 void
@@ -1092,6 +1150,30 @@ UnicodeStringTest::TestMiscellaneous()
     if(test1.hasMetaData() || UnicodeString().hasMetaData()) {
         errln("UnicodeString::hasMetaData() returns TRUE");
     }
+
+    // test getTerminatedBuffer() on a truncated, shared, heap-allocated string
+    test1=UNICODE_STRING_SIMPLE("abcdefghijklmnopqrstuvwxyz0123456789.");
+    test1.truncate(36);  // ensure length()<getCapacity()
+    test2=test1;  // share the buffer
+    test1.truncate(5);
+    if(test1.length()!=5 || test1.getTerminatedBuffer()[5]!=0) {
+        errln("UnicodeString(shared buffer).truncate() failed");
+    }
+    if(test2.length()!=36 || test2[5]!=0x66 || u_strlen(test2.getTerminatedBuffer())!=36) {
+        errln("UnicodeString(shared buffer).truncate().getTerminatedBuffer() "
+              "modified another copy of the string!");
+    }
+    test1=UNICODE_STRING_SIMPLE("abcdefghijklmnopqrstuvwxyz0123456789.");
+    test1.truncate(36);  // ensure length()<getCapacity()
+    test2=test1;  // share the buffer
+    test1.remove();
+    if(test1.length()!=0 || test1.getTerminatedBuffer()[0]!=0) {
+        errln("UnicodeString(shared buffer).remove() failed");
+    }
+    if(test2.length()!=36 || test2[0]!=0x61 || u_strlen(test2.getTerminatedBuffer())!=36) {
+        errln("UnicodeString(shared buffer).remove().getTerminatedBuffer() "
+              "modified another copy of the string!");
+    }
 }
 
 void
@@ -1201,7 +1283,7 @@ UnicodeStringTest::TestStackAllocation()
 
     // test the UChar32 constructor
     UnicodeString c32Test((UChar32)0x10ff2a);
-    if( c32Test.length() != UTF_CHAR_LENGTH(0x10ff2a) ||
+    if( c32Test.length() != U16_LENGTH(0x10ff2a) ||
         c32Test.char32At(c32Test.length() - 1) != 0x10ff2a
     ) {
         errln("The UnicodeString(UChar32) constructor does not work with a 0x10ff2a filler");
@@ -1209,7 +1291,7 @@ UnicodeStringTest::TestStackAllocation()
 
     // test the (new) capacity constructor
     UnicodeString capTest(5, (UChar32)0x2a, 5);
-    if( capTest.length() != 5 * UTF_CHAR_LENGTH(0x2a) ||
+    if( capTest.length() != 5 * U16_LENGTH(0x2a) ||
         capTest.char32At(0) != 0x2a ||
         capTest.char32At(4) != 0x2a
     ) {
@@ -1217,7 +1299,7 @@ UnicodeStringTest::TestStackAllocation()
     }
 
     capTest = UnicodeString(5, (UChar32)0x10ff2a, 5);
-    if( capTest.length() != 5 * UTF_CHAR_LENGTH(0x10ff2a) ||
+    if( capTest.length() != 5 * U16_LENGTH(0x10ff2a) ||
         capTest.char32At(0) != 0x10ff2a ||
         capTest.char32At(4) != 0x10ff2a
     ) {
@@ -1234,7 +1316,7 @@ UnicodeStringTest::TestStackAllocation()
  * Test the unescape() function.
  */
 void UnicodeStringTest::TestUnescape(void) {
-    UnicodeString IN("abc\\u4567 \\n\\r \\U00101234xyz\\x1\\x{5289}\\x1b");
+    UnicodeString IN("abc\\u4567 \\n\\r \\U00101234xyz\\x1\\x{5289}\\x1b", -1, US_INV);
     UnicodeString OUT("abc");
     OUT.append((UChar)0x4567);
     OUT.append(" ");
@@ -1373,6 +1455,18 @@ UnicodeStringTest::TestBogus() {
     if(test3.getBuffer()!=0 || test3.getBuffer(20)!=0 || test3.getTerminatedBuffer()!=0) {
         errln("bogus.getBuffer()!=0");
     }
+    if (test1.indexOf(test3) != -1) {
+        errln("bogus.indexOf() != -1");
+    }
+    if (test1.lastIndexOf(test3) != -1) {
+        errln("bogus.lastIndexOf() != -1");
+    }
+    if (test1.caseCompare(test3, U_FOLD_CASE_DEFAULT) != 1 || test3.caseCompare(test1, U_FOLD_CASE_DEFAULT) != -1) {
+        errln("caseCompare() doesn't work with bogus strings");
+    }
+    if (test1.compareCodePointOrder(test3) != 1 || test3.compareCodePointOrder(test1) != -1) {
+        errln("compareCodePointOrder() doesn't work with bogus strings");
+    }
 
     // verify that non-assignment modifications fail and do not revive a bogus string
     test3.setToBogus();
@@ -1636,15 +1730,382 @@ UnicodeStringTest::TestStringEnumeration() {
     if(ten.clone()!=NULL) {
         errln("StringEnumeration.clone()!=NULL");
     }
+
+    // test that uenum_openFromStringEnumeration() works
+    // Need a heap allocated string enumeration because it is adopted by the UEnumeration.
+    StringEnumeration *newTen = new TestEnumeration;
+    status=U_ZERO_ERROR;
+    UEnumeration *uten = uenum_openFromStringEnumeration(newTen, &status);
+    if (uten==NULL || U_FAILURE(status)) {
+        errln("fail at file %s, line %d, UErrorCode is %s\n", __FILE__, __LINE__, u_errorName(status));
+        return;
+    }
+    
+    // test  uenum_next()
+    for(i=0; i<LENGTHOF(testEnumStrings); ++i) {
+        status=U_ZERO_ERROR;
+        pc=uenum_next(uten, &length, &status);
+        if(U_FAILURE(status) || pc==NULL || strcmp(pc, testEnumStrings[i]) != 0) {
+            errln("File %s, line %d, StringEnumeration.next(%d) failed", __FILE__, __LINE__, i);
+        }
+    }
+    status=U_ZERO_ERROR;
+    if(uenum_next(uten, &length, &status)!=NULL) {
+        errln("File %s, line %d, uenum_next(done)!=NULL");
+    }
+
+    // test the uenum_unext()
+    uenum_reset(uten, &status);
+    for(i=0; i<LENGTHOF(testEnumStrings); ++i) {
+        status=U_ZERO_ERROR;
+        pu=uenum_unext(uten, &length, &status);
+        s=UnicodeString(testEnumStrings[i], "");
+        if(U_FAILURE(status) || pu==NULL || length!=s.length() || UnicodeString(TRUE, pu, length)!=s) {
+            errln("File %s, Line %d, uenum_unext(%d) failed", __FILE__, __LINE__, i);
+        }
+    }
+    status=U_ZERO_ERROR;
+    if(uenum_unext(uten, &length, &status)!=NULL) {
+        errln("File %s, Line %d, uenum_unext(done)!=NULL" __FILE__, __LINE__);
+    }
+
+    uenum_close(uten);
+}
+
+/*
+ * Namespace test, to make sure that macros like UNICODE_STRING include the
+ * namespace qualifier.
+ *
+ * Define a (bogus) UnicodeString class in another namespace and check for ambiguity.
+ */
+namespace bogus {
+    class UnicodeString {
+    public:
+        enum EInvariant { kInvariant };
+        UnicodeString() : i(1) {}
+        UnicodeString(UBool /*isTerminated*/, const UChar * /*text*/, int32_t textLength) : i(textLength) {}
+        UnicodeString(const char * /*src*/, int32_t length, enum EInvariant /*inv*/
+) : i(length) {}
+    private:
+        int32_t i;
+    };
+}
+
+void
+UnicodeStringTest::TestNameSpace() {
+    // Provoke name collision unless the UnicodeString macros properly
+    // qualify the icu::UnicodeString class.
+    using namespace bogus;
+
+    // Use all UnicodeString macros from unistr.h.
+    icu::UnicodeString s1=icu::UnicodeString("abc", 3, US_INV);
+    icu::UnicodeString s2=UNICODE_STRING("def", 3);
+    icu::UnicodeString s3=UNICODE_STRING_SIMPLE("ghi");
+
+    // Make sure the compiler does not optimize away instantiation of s1, s2, s3.
+    icu::UnicodeString s4=s1+s2+s3;
+    if(s4.length()!=9) {
+        errln("Something wrong with UnicodeString::operator+().");
+    }
+}
+
+void
+UnicodeStringTest::TestUTF32() {
+    // Input string length US_STACKBUF_SIZE to cause overflow of the
+    // initially chosen fStackBuffer due to supplementary characters.
+    static const UChar32 utf32[] = {
+        0x41, 0xd900, 0x61, 0xdc00, -1, 0x110000, 0x5a, 0x50000, 0x7a,
+        0x10000, 0x20000, 0xe0000, 0x10ffff
+    };
+    static const UChar expected_utf16[] = {
+        0x41, 0xfffd, 0x61, 0xfffd, 0xfffd, 0xfffd, 0x5a, 0xd900, 0xdc00, 0x7a,
+        0xd800, 0xdc00, 0xd840, 0xdc00, 0xdb40, 0xdc00, 0xdbff, 0xdfff
+    };
+    UnicodeString from32 = UnicodeString::fromUTF32(utf32, LENGTHOF(utf32));
+    UnicodeString expected(FALSE, expected_utf16, LENGTHOF(expected_utf16));
+    if(from32 != expected) {
+        errln("UnicodeString::fromUTF32() did not create the expected string.");
+    }
+
+    static const UChar utf16[] = {
+        0x41, 0xd900, 0x61, 0xdc00, 0x5a, 0xd900, 0xdc00, 0x7a, 0xd800, 0xdc00, 0xdbff, 0xdfff
+    };
+    static const UChar32 expected_utf32[] = {
+        0x41, 0xfffd, 0x61, 0xfffd, 0x5a, 0x50000, 0x7a, 0x10000, 0x10ffff
+    };
+    UChar32 result32[16];
+    UErrorCode errorCode = U_ZERO_ERROR;
+    int32_t length32 =
+        UnicodeString(FALSE, utf16, LENGTHOF(utf16)).
+        toUTF32(result32, LENGTHOF(result32), errorCode);
+    if( length32 != LENGTHOF(expected_utf32) ||
+        0 != uprv_memcmp(result32, expected_utf32, length32*4) ||
+        result32[length32] != 0
+    ) {
+        errln("UnicodeString::toUTF32() did not create the expected string.");
+    }
+}
+
+class TestCheckedArrayByteSink : public CheckedArrayByteSink {
+public:
+    TestCheckedArrayByteSink(char* outbuf, int32_t capacity)
+            : CheckedArrayByteSink(outbuf, capacity), calledFlush(FALSE) {}
+    virtual void Flush() { calledFlush = TRUE; }
+    UBool calledFlush;
+};
+
+void
+UnicodeStringTest::TestUTF8() {
+    static const uint8_t utf8[] = {
+        // Code points:
+        // 0x41, 0xd900,
+        // 0x61, 0xdc00,
+        // 0x110000, 0x5a,
+        // 0x50000, 0x7a,
+        // 0x10000, 0x20000,
+        // 0xe0000, 0x10ffff
+        0x41, 0xed, 0xa4, 0x80,
+        0x61, 0xed, 0xb0, 0x80,
+        0xf4, 0x90, 0x80, 0x80, 0x5a,
+        0xf1, 0x90, 0x80, 0x80, 0x7a,
+        0xf0, 0x90, 0x80, 0x80, 0xf0, 0xa0, 0x80, 0x80,
+        0xf3, 0xa0, 0x80, 0x80, 0xf4, 0x8f, 0xbf, 0xbf
+    };
+    static const UChar expected_utf16[] = {
+        0x41, 0xfffd,
+        0x61, 0xfffd,
+        0xfffd, 0x5a,
+        0xd900, 0xdc00, 0x7a,
+        0xd800, 0xdc00, 0xd840, 0xdc00,
+        0xdb40, 0xdc00, 0xdbff, 0xdfff
+    };
+    UnicodeString from8 = UnicodeString::fromUTF8(StringPiece((const char *)utf8, (int32_t)sizeof(utf8)));
+    UnicodeString expected(FALSE, expected_utf16, LENGTHOF(expected_utf16));
+
+    if(from8 != expected) {
+        errln("UnicodeString::fromUTF8(StringPiece) did not create the expected string.");
+    }
+#if U_HAVE_STD_STRING
+    std::string utf8_string((const char *)utf8, sizeof(utf8));
+    UnicodeString from8b = UnicodeString::fromUTF8(utf8_string);
+    if(from8b != expected) {
+        errln("UnicodeString::fromUTF8(std::string) did not create the expected string.");
+    }
+#endif
+
+    static const UChar utf16[] = {
+        0x41, 0xd900, 0x61, 0xdc00, 0x5a, 0xd900, 0xdc00, 0x7a, 0xd800, 0xdc00, 0xdbff, 0xdfff
+    };
+    static const uint8_t expected_utf8[] = {
+        0x41, 0xef, 0xbf, 0xbd, 0x61, 0xef, 0xbf, 0xbd, 0x5a, 0xf1, 0x90, 0x80, 0x80, 0x7a,
+        0xf0, 0x90, 0x80, 0x80, 0xf4, 0x8f, 0xbf, 0xbf
+    };
+    UnicodeString us(FALSE, utf16, LENGTHOF(utf16));
+
+    char buffer[64];
+    TestCheckedArrayByteSink sink(buffer, (int32_t)sizeof(buffer));
+    us.toUTF8(sink);
+    if( sink.NumberOfBytesWritten() != (int32_t)sizeof(expected_utf8) ||
+        0 != uprv_memcmp(buffer, expected_utf8, sizeof(expected_utf8))
+    ) {
+        errln("UnicodeString::toUTF8() did not create the expected string.");
+    }
+    if(!sink.calledFlush) {
+        errln("UnicodeString::toUTF8(sink) did not sink.Flush().");
+    }
+#if U_HAVE_STD_STRING
+    // Initial contents for testing that toUTF8String() appends.
+    std::string result8 = "-->";
+    std::string expected8 = "-->" + std::string((const char *)expected_utf8, sizeof(expected_utf8));
+    // Use the return value just for testing.
+    std::string &result8r = us.toUTF8String(result8);
+    if(result8r != expected8 || &result8r != &result8) {
+        errln("UnicodeString::toUTF8String() did not create the expected string.");
+    }
+#endif
+}
+
+// Test if this compiler supports Return Value Optimization of unnamed temporary objects.
+static UnicodeString wrapUChars(const UChar *uchars) {
+    return UnicodeString(TRUE, uchars, -1);
+}
+
+void
+UnicodeStringTest::TestReadOnlyAlias() {
+    UChar uchars[]={ 0x61, 0x62, 0 };
+    UnicodeString alias(TRUE, uchars, 2);
+    if(alias.length()!=2 || alias.getBuffer()!=uchars || alias.getTerminatedBuffer()!=uchars) {
+        errln("UnicodeString read-only-aliasing constructor does not behave as expected.");
+        return;
+    }
+    alias.truncate(1);
+    if(alias.length()!=1 || alias.getBuffer()!=uchars) {
+        errln("UnicodeString(read-only-alias).truncate() did not preserve aliasing as expected.");
+    }
+    if(alias.getTerminatedBuffer()==uchars) {
+        errln("UnicodeString(read-only-alias).truncate().getTerminatedBuffer() "
+              "did not allocate and copy as expected.");
+    }
+    if(uchars[1]!=0x62) {
+        errln("UnicodeString(read-only-alias).truncate().getTerminatedBuffer() "
+              "modified the original buffer.");
+    }
+    if(1!=u_strlen(alias.getTerminatedBuffer())) {
+        errln("UnicodeString(read-only-alias).truncate().getTerminatedBuffer() "
+              "does not return a buffer terminated at the proper length.");
+    }
+
+    alias.setTo(TRUE, uchars, 2);
+    if(alias.length()!=2 || alias.getBuffer()!=uchars || alias.getTerminatedBuffer()!=uchars) {
+        errln("UnicodeString read-only-aliasing setTo() does not behave as expected.");
+        return;
+    }
+    alias.remove();
+    if(alias.length()!=0) {
+        errln("UnicodeString(read-only-alias).remove() did not work.");
+    }
+    if(alias.getTerminatedBuffer()==uchars) {
+        errln("UnicodeString(read-only-alias).remove().getTerminatedBuffer() "
+              "did not un-alias as expected.");
+    }
+    if(uchars[0]!=0x61) {
+        errln("UnicodeString(read-only-alias).remove().getTerminatedBuffer() "
+              "modified the original buffer.");
+    }
+    if(0!=u_strlen(alias.getTerminatedBuffer())) {
+        errln("UnicodeString.setTo(read-only-alias).remove().getTerminatedBuffer() "
+              "does not return a buffer terminated at length 0.");
+    }
+
+    UnicodeString longString=UNICODE_STRING_SIMPLE("abcdefghijklmnopqrstuvwxyz0123456789");
+    alias.setTo(FALSE, longString.getBuffer(), longString.length());
+    alias.remove(0, 10);
+    if(longString.compare(10, INT32_MAX, alias)!=0 || alias.getBuffer()!=longString.getBuffer()+10) {
+        errln("UnicodeString.setTo(read-only-alias).remove(0, 10) did not preserve aliasing as expected.");
+    }
+    alias.setTo(FALSE, longString.getBuffer(), longString.length());
+    alias.remove(27, 99);
+    if(longString.compare(0, 27, alias)!=0 || alias.getBuffer()!=longString.getBuffer()) {
+        errln("UnicodeString.setTo(read-only-alias).remove(27, 99) did not preserve aliasing as expected.");
+    }
+    alias.setTo(FALSE, longString.getBuffer(), longString.length());
+    alias.retainBetween(6, 30);
+    if(longString.compare(6, 24, alias)!=0 || alias.getBuffer()!=longString.getBuffer()+6) {
+        errln("UnicodeString.setTo(read-only-alias).retainBetween(6, 30) did not preserve aliasing as expected.");
+    }
+
+    UChar abc[]={ 0x61, 0x62, 0x63, 0 };
+    UBool hasRVO= wrapUChars(abc).getBuffer()==abc;
+
+    UnicodeString temp;
+    temp.fastCopyFrom(longString.tempSubString());
+    if(temp!=longString || (hasRVO && temp.getBuffer()!=longString.getBuffer())) {
+        errln("UnicodeString.tempSubString() failed");
+    }
+    temp.fastCopyFrom(longString.tempSubString(-3, 5));
+    if(longString.compare(0, 5, temp)!=0 || (hasRVO && temp.getBuffer()!=longString.getBuffer())) {
+        errln("UnicodeString.tempSubString(-3, 5) failed");
+    }
+    temp.fastCopyFrom(longString.tempSubString(17));
+    if(longString.compare(17, INT32_MAX, temp)!=0 || (hasRVO && temp.getBuffer()!=longString.getBuffer()+17)) {
+        errln("UnicodeString.tempSubString(17) failed");
+    }
+    temp.fastCopyFrom(longString.tempSubString(99));
+    if(!temp.isEmpty()) {
+        errln("UnicodeString.tempSubString(99) failed");
+    }
+    temp.fastCopyFrom(longString.tempSubStringBetween(6));
+    if(longString.compare(6, INT32_MAX, temp)!=0 || (hasRVO && temp.getBuffer()!=longString.getBuffer()+6)) {
+        errln("UnicodeString.tempSubStringBetween(6) failed");
+    }
+    temp.fastCopyFrom(longString.tempSubStringBetween(8, 18));
+    if(longString.compare(8, 10, temp)!=0 || (hasRVO && temp.getBuffer()!=longString.getBuffer()+8)) {
+        errln("UnicodeString.tempSubStringBetween(8, 18) failed");
+    }
+    UnicodeString bogusString;
+    bogusString.setToBogus();
+    temp.fastCopyFrom(bogusString.tempSubStringBetween(8, 18));
+    if(!temp.isBogus()) {
+        errln("UnicodeString.setToBogus().tempSubStringBetween(8, 18) failed");
+    }
 }
 
 void
-UnicodeStringTest::TestCharString() {
-    static const char originalCStr[] =
-        "This is a large string that is meant to over flow the internal buffer of CharString. At the time of writing this test, the internal buffer is 128 bytes.";
-    CharString chStr(originalCStr);
-    if (strcmp(originalCStr, chStr) != 0) {
-        errln("CharString doesn't work with large strings.");
+UnicodeStringTest::doTestAppendable(UnicodeString &dest, Appendable &app) {
+    static const UChar cde[3]={ 0x63, 0x64, 0x65 };
+    static const UChar fg[3]={ 0x66, 0x67, 0 };
+    if(!app.reserveAppendCapacity(12)) {
+        errln("Appendable.reserve(12) failed");
+    }
+    app.appendCodeUnit(0x61);
+    app.appendCodePoint(0x62);
+    app.appendCodePoint(0x50000);
+    app.appendString(cde, 3);
+    app.appendString(fg, -1);
+    UChar scratch[3];
+    int32_t capacity=-1;
+    UChar *buffer=app.getAppendBuffer(3, 3, scratch, 3, &capacity);
+    if(capacity<3) {
+        errln("Appendable.getAppendBuffer(min=3) returned capacity=%d<3", (int)capacity);
+        return;
+    }
+    static const UChar hij[3]={ 0x68, 0x69, 0x6a };
+    u_memcpy(buffer, hij, 3);
+    app.appendString(buffer, 3);
+    if(dest!=UNICODE_STRING_SIMPLE("ab\\U00050000cdefghij").unescape()) {
+        errln("Appendable.append(...) failed");
+    }
+    buffer=app.getAppendBuffer(0, 3, scratch, 3, &capacity);
+    if(buffer!=NULL || capacity!=0) {
+        errln("Appendable.getAppendBuffer(min=0) failed");
+    }
+    capacity=1;
+    buffer=app.getAppendBuffer(3, 3, scratch, 2, &capacity);
+    if(buffer!=NULL || capacity!=0) {
+        errln("Appendable.getAppendBuffer(scratch<min) failed");
     }
 }
 
+class SimpleAppendable : public Appendable {
+public:
+    explicit SimpleAppendable(UnicodeString &dest) : str(dest) {}
+    virtual UBool appendCodeUnit(UChar c) { str.append(c); return TRUE; }
+    SimpleAppendable &reset() { str.remove(); return *this; }
+private:
+    UnicodeString &str;
+};
+
+void
+UnicodeStringTest::TestAppendable() {
+    UnicodeString dest;
+    SimpleAppendable app(dest);
+    doTestAppendable(dest, app);
+}
+
+void
+UnicodeStringTest::TestUnicodeStringImplementsAppendable() {
+    UnicodeString dest;
+    UnicodeStringAppendable app(dest);
+    doTestAppendable(dest, app);
+}
+
+void
+UnicodeStringTest::TestSizeofUnicodeString() {
+    // See the comments in unistr.h near the declaration of UnicodeString's fields.
+    size_t sizeofUniStr=sizeof(UnicodeString);
+    size_t expected;
+    switch(sizeof(void *)) {
+    case 4:
+        expected=32;
+        break;
+    case 8:
+        expected=40;
+        break;
+    default:
+        logln("This platform has neither 32-bit nor 64-bit pointers.");
+        return;
+    }
+    if(expected!=sizeofUniStr) {
+        errln("sizeof(UnicodeString)=%d, expected %d", (int)sizeofUniStr, (int)expected);
+    }
+}