]> git.saurik.com Git - wxWidgets.git/blobdiff - tests/strings/unicode.cpp
Somehow, setting a tint color makes gauge work :/.
[wxWidgets.git] / tests / strings / unicode.cpp
index 94f4c90e8cac408b158c409db50a3c3d40ebb63f..29b7777a2a9375517991376df50f5731d93c06f3 100644 (file)
@@ -3,7 +3,6 @@
 // Purpose:     Unicode unit test
 // Author:      Vadim Zeitlin, Wlodzimierz ABX Skiba
 // Created:     2004-04-28
 // Purpose:     Unicode unit test
 // Author:      Vadim Zeitlin, Wlodzimierz ABX Skiba
 // Created:     2004-04-28
-// RCS-ID:      $Id$
 // Copyright:   (c) 2004 Vadim Zeitlin, Wlodzimierz Skiba
 ///////////////////////////////////////////////////////////////////////////////
 
 // Copyright:   (c) 2004 Vadim Zeitlin, Wlodzimierz Skiba
 ///////////////////////////////////////////////////////////////////////////////
 
 #endif
 
 #ifndef WX_PRECOMP
 #endif
 
 #ifndef WX_PRECOMP
+    #include "wx/wx.h"
 #endif // WX_PRECOMP
 
 #endif // WX_PRECOMP
 
+#include "wx/encconv.h"
+
+// ----------------------------------------------------------------------------
+// helper class holding the matching MB and WC strings
+// ----------------------------------------------------------------------------
+
+struct StringConversionData
+{
+    // either str or wcs (but not both) may be NULL, this means that the conversion
+    // to it should fail
+    StringConversionData(const char *str_, const wchar_t *wcs_, int flags_ = 0)
+        : str(str_), wcs(wcs_), flags(flags_)
+    {
+    }
+
+    const char * const str;
+    const wchar_t * const wcs;
+
+    enum
+    {
+        TEST_BOTH  = 0, // test both str -> wcs and wcs -> str
+        ONLY_MB2WC = 1  // only test str -> wcs conversion
+    };
+
+    const int flags;
+
+    // test that the conversion between str and wcs (subject to flags) succeeds
+    //
+    // the first argument is the index in the test array and is used solely for
+    // diagnostics
+    void Test(size_t n, wxMBConv& conv) const
+    {
+        if ( str )
+        {
+            wxWCharBuffer wbuf = conv.cMB2WC(str);
+
+            if ( wcs )
+            {
+                CPPUNIT_ASSERT_MESSAGE
+                (
+                    Message(n, "MB2WC failed"),
+                    wbuf.data()
+                );
+
+                CPPUNIT_ASSERT_MESSAGE
+                (
+                    Message(n, "MB2WC", wbuf, wcs),
+                    wxStrcmp(wbuf, wcs) == 0
+                );
+            }
+            else // conversion is supposed to fail
+            {
+                CPPUNIT_ASSERT_MESSAGE
+                (
+                    Message(n, "MB2WC succeeded"),
+                    !wbuf.data()
+                );
+            }
+        }
+
+        if ( wcs && !(flags & ONLY_MB2WC) )
+        {
+            wxCharBuffer buf = conv.cWC2MB(wcs);
+
+            if ( str )
+            {
+                CPPUNIT_ASSERT_MESSAGE
+                (
+                    Message(n, "WC2MB failed"),
+                    buf.data()
+                );
+
+                CPPUNIT_ASSERT_MESSAGE
+                (
+                    Message(n, "WC2MB", buf, str),
+                    strcmp(buf, str) == 0
+                );
+            }
+            else
+            {
+                CPPUNIT_ASSERT_MESSAGE
+                (
+                    Message(n, "WC2MB succeeded"),
+                    !buf.data()
+                );
+            }
+        }
+    }
+
+private:
+    static std::string
+    Message(size_t n, const wxString& msg)
+    {
+        return std::string(wxString::Format("#%lu: %s", (unsigned long)n, msg));
+    }
+
+    template <typename T>
+    static std::string
+    Message(size_t n,
+            const char *func,
+            const wxCharTypeBuffer<T>& actual,
+            const T *expected)
+    {
+        return Message(n,
+                       wxString::Format("%s returned \"%s\", expected \"%s\"",
+                                        func, actual.data(), expected));
+    }
+};
+
 // ----------------------------------------------------------------------------
 // test class
 // ----------------------------------------------------------------------------
 // ----------------------------------------------------------------------------
 // test class
 // ----------------------------------------------------------------------------
@@ -32,28 +141,31 @@ public:
 private:
     CPPUNIT_TEST_SUITE( UnicodeTestCase );
         CPPUNIT_TEST( ToFromAscii );
 private:
     CPPUNIT_TEST_SUITE( UnicodeTestCase );
         CPPUNIT_TEST( ToFromAscii );
-#if wxUSE_WCHAR_T
         CPPUNIT_TEST( ConstructorsWithConversion );
         CPPUNIT_TEST( ConstructorsWithConversion );
-        CPPUNIT_TEST( Conversion );
+        CPPUNIT_TEST( ConversionFixed );
+        CPPUNIT_TEST( ConversionWithNULs );
         CPPUNIT_TEST( ConversionUTF7 );
         CPPUNIT_TEST( ConversionUTF8 );
         CPPUNIT_TEST( ConversionUTF7 );
         CPPUNIT_TEST( ConversionUTF8 );
-#endif // wxUSE_WCHAR_T
+        CPPUNIT_TEST( ConversionUTF16 );
+        CPPUNIT_TEST( ConversionUTF32 );
+        CPPUNIT_TEST( IsConvOk );
+#if wxUSE_UNICODE
+        CPPUNIT_TEST( Iteration );
+#endif
     CPPUNIT_TEST_SUITE_END();
 
     void ToFromAscii();
     CPPUNIT_TEST_SUITE_END();
 
     void ToFromAscii();
-#if wxUSE_WCHAR_T
     void ConstructorsWithConversion();
     void ConstructorsWithConversion();
-    void Conversion();
+    void ConversionFixed();
+    void ConversionWithNULs();
     void ConversionUTF7();
     void ConversionUTF8();
     void ConversionUTF7();
     void ConversionUTF8();
-
-    // test if converting s using the given encoding gives ws and vice versa
-    //
-    // if either of the first 2 arguments is NULL, the conversion is supposed
-    // to fail
-    void DoTestConversion(const char *s, const wchar_t *w, wxCSConv& conv);
-#endif // wxUSE_WCHAR_T
-
+    void ConversionUTF16();
+    void ConversionUTF32();
+    void IsConvOk();
+#if wxUSE_UNICODE
+    void Iteration();
+#endif
 
     DECLARE_NO_COPY_CLASS(UnicodeTestCase)
 };
 
     DECLARE_NO_COPY_CLASS(UnicodeTestCase)
 };
@@ -61,7 +173,7 @@ private:
 // register in the unnamed registry so that these tests are run by default
 CPPUNIT_TEST_SUITE_REGISTRATION( UnicodeTestCase );
 
 // register in the unnamed registry so that these tests are run by default
 CPPUNIT_TEST_SUITE_REGISTRATION( UnicodeTestCase );
 
-// also include in it's own registry so that these tests can be run alone
+// also include in its own registry so that these tests can be run alone
 CPPUNIT_TEST_SUITE_NAMED_REGISTRATION( UnicodeTestCase, "UnicodeTestCase" );
 
 UnicodeTestCase::UnicodeTestCase()
 CPPUNIT_TEST_SUITE_NAMED_REGISTRATION( UnicodeTestCase, "UnicodeTestCase" );
 
 UnicodeTestCase::UnicodeTestCase()
@@ -82,162 +194,283 @@ void UnicodeTestCase::ToFromAscii()
     TEST_TO_FROM_ASCII( "additional \" special \t test \\ component \n :-)" );
 }
 
     TEST_TO_FROM_ASCII( "additional \" special \t test \\ component \n :-)" );
 }
 
-#if wxUSE_WCHAR_T
 void UnicodeTestCase::ConstructorsWithConversion()
 {
 void UnicodeTestCase::ConstructorsWithConversion()
 {
-    // the string "Déjà" in UTF-8 and wchar_t:
+    // the string "Déjà" in UTF-8 and wchar_t:
     const unsigned char utf8Buf[] = {0x44,0xC3,0xA9,0x6A,0xC3,0xA0,0};
     const unsigned char utf8Buf[] = {0x44,0xC3,0xA9,0x6A,0xC3,0xA0,0};
-    const wchar_t wchar[] = {0x44,0xE9,0x6A,0xE0,0};
-    const unsigned char utf8subBuf[] = {0x44,0xC3,0xA9,0x6A,0}; // just "Déj"
+    const unsigned char utf8subBuf[] = {0x44,0xC3,0xA9,0x6A,0}; // just "Déj"
     const char *utf8 = (char *)utf8Buf;
     const char *utf8sub = (char *)utf8subBuf;
 
     wxString s1(utf8, wxConvUTF8);
     const char *utf8 = (char *)utf8Buf;
     const char *utf8sub = (char *)utf8subBuf;
 
     wxString s1(utf8, wxConvUTF8);
-    wxString s2(wchar, wxConvUTF8);
 
 #if wxUSE_UNICODE
 
 #if wxUSE_UNICODE
-    CPPUNIT_ASSERT( s1 == wchar );
-    CPPUNIT_ASSERT( s2 == wchar );
+    const wchar_t wchar[] = {0x44,0xE9,0x6A,0xE0,0};
+    CPPUNIT_ASSERT_EQUAL( wchar, s1 );
+
+    wxString s2(wchar);
+    CPPUNIT_ASSERT_EQUAL( wchar, s2 );
+    CPPUNIT_ASSERT_EQUAL( wxString::FromUTF8(utf8), s2 );
 #else
 #else
-    CPPUNIT_ASSERT( s1 == utf8 );
-    CPPUNIT_ASSERT( s2 == utf8 );
+    CPPUNIT_ASSERT_EQUAL( utf8, s1 );
 #endif
 
     wxString sub(utf8sub, wxConvUTF8); // "Dej" substring
     wxString s3(utf8, wxConvUTF8, 4);
 #endif
 
     wxString sub(utf8sub, wxConvUTF8); // "Dej" substring
     wxString s3(utf8, wxConvUTF8, 4);
+    CPPUNIT_ASSERT_EQUAL( sub, s3 );
+
+#if wxUSE_UNICODE
     wxString s4(wchar, wxConvUTF8, 3);
     wxString s4(wchar, wxConvUTF8, 3);
+    CPPUNIT_ASSERT_EQUAL( sub, s4 );
+
+    // conversion should stop with failure at pos 35
+    wxString s("\t[pl]open.format.Sformatuj dyskietk\xea=gfloppy %f", wxConvUTF8);
+    CPPUNIT_ASSERT( s.empty() );
+#endif // wxUSE_UNICODE
+
+
+    // test using Unicode strings together with char* strings (this must work
+    // in ANSI mode as well, of course):
+    wxString s5("ascii");
+    CPPUNIT_ASSERT_EQUAL( "ascii", s5 );
 
 
-    CPPUNIT_ASSERT( s3 == sub );
-    CPPUNIT_ASSERT( s4 == sub );
+    s5 += " value";
+
+    CPPUNIT_ASSERT( strcmp(s5.mb_str(), "ascii value") == 0 );
+    CPPUNIT_ASSERT_EQUAL( "ascii value", s5 );
+    CPPUNIT_ASSERT( s5 != "SomethingElse" );
+}
+
+void UnicodeTestCase::ConversionFixed()
+{
+    size_t len;
 
 #if wxUSE_UNICODE
 
 #if wxUSE_UNICODE
-    CPPUNIT_ASSERT ( wxString("\t[pl]open.format.Sformatuj dyskietkê=gfloppy %f", 
-                               wxConvUTF8) == wxT("") ); //should stop at pos 35 
-#endif
+    wxConvLibc.cWC2MB(L"", 0, &len);
+#else // !wxUSE_UNICODE
+    wxConvLibc.cMB2WC("", 0, &len);
+#endif // wxUSE_UNICODE/!wxUSE_UNICODE
+
+    CPPUNIT_ASSERT_EQUAL( 0, len );
+
+#if wxUSE_UNICODE
+    // check that when we convert a fixed number of characters we obtain the
+    // expected return value
+    CPPUNIT_ASSERT_EQUAL( 0, wxConvLibc.ToWChar(NULL, 0, "", 0) );
+    CPPUNIT_ASSERT_EQUAL( 1, wxConvLibc.ToWChar(NULL, 0, "x", 1) );
+    CPPUNIT_ASSERT_EQUAL( 2, wxConvLibc.ToWChar(NULL, 0, "x", 2) );
+    CPPUNIT_ASSERT_EQUAL( 2, wxConvLibc.ToWChar(NULL, 0, "xy", 2) );
+#endif // wxUSE_UNICODE
 }
 
 }
 
-void UnicodeTestCase::Conversion()
+void UnicodeTestCase::ConversionWithNULs()
 {
 #if wxUSE_UNICODE
 {
 #if wxUSE_UNICODE
-        wxString szTheString(L"The\0String", wxConvLibc, 10);
-        wxCharBuffer theBuffer = szTheString.mb_str();
+    static const size_t lenNulString = 10;
 
 
-        CPPUNIT_ASSERT( memcmp(theBuffer.data(), "The\0String", 11) == 0 );
+    wxString szTheString(L"The\0String", wxConvLibc, lenNulString);
+    wxCharBuffer theBuffer = szTheString.mb_str();
 
 
-        wxString szTheString2("The\0String", wxConvLocal, 10);
-        CPPUNIT_ASSERT( szTheString2.length() == 11 );
-        CPPUNIT_ASSERT( wxTmemcmp(szTheString2.c_str(), L"The\0String", 11) == 0 );
-#else
-        wxString szTheString(wxT("TheString"));
-        szTheString.insert(3, 1, '\0');
-        wxWCharBuffer theBuffer = szTheString.wc_str(wxConvLibc);
+    CPPUNIT_ASSERT( memcmp(theBuffer.data(), "The\0String",
+                    lenNulString + 1) == 0 );
 
 
-        CPPUNIT_ASSERT( memcmp(theBuffer.data(), L"The\0String", 11 * sizeof(wchar_t)) == 0 );
+    wxString szTheString2("The\0String", wxConvLocal, lenNulString);
+    CPPUNIT_ASSERT_EQUAL( lenNulString, szTheString2.length() );
+    CPPUNIT_ASSERT( wxTmemcmp(szTheString2.c_str(), L"The\0String",
+                    lenNulString + 1) == 0 );
+#else // !wxUSE_UNICODE
+    wxString szTheString("TheString");
+    szTheString.insert(3, 1, '\0');
+    wxWCharBuffer theBuffer = szTheString.wc_str(wxConvLibc);
 
 
-        wxString szLocalTheString(wxT("TheString"));
-        szLocalTheString.insert(3, 1, '\0');
-        wxWCharBuffer theLocalBuffer = szLocalTheString.wc_str(wxConvLocal);
+    CPPUNIT_ASSERT( memcmp(theBuffer.data(), L"The\0String", 11 * sizeof(wchar_t)) == 0 );
 
 
-        CPPUNIT_ASSERT( memcmp(theLocalBuffer.data(), L"The\0String", 11 * sizeof(wchar_t)) == 0 );
-#endif
+    wxString szLocalTheString("TheString");
+    szLocalTheString.insert(3, 1, '\0');
+    wxWCharBuffer theLocalBuffer = szLocalTheString.wc_str(wxConvLocal);
+
+    CPPUNIT_ASSERT( memcmp(theLocalBuffer.data(), L"The\0String", 11 * sizeof(wchar_t)) == 0 );
+#endif // wxUSE_UNICODE/!wxUSE_UNICODE
 }
 
 }
 
-#if !wxUSE_UNICODE
-// in case wcscmp is missing
-//
-static int wx_wcscmp(const wchar_t *s1, const wchar_t *s2)
+void UnicodeTestCase::ConversionUTF7()
 {
 {
-    while (*s1 == *s2 && *s1 != 0)
+    static const StringConversionData utf7data[] =
+    {
+        // normal fragments
+        StringConversionData("+AKM-", L"\xa3"),
+        StringConversionData("+AOk-t+AOk-", L"\xe9t\xe9"),
+
+        // this one is an alternative valid encoding of the same string
+        StringConversionData("+AOk-t+AOk", L"\xe9t\xe9",
+                             StringConversionData::ONLY_MB2WC),
+
+        // some special cases
+        StringConversionData("+-", L"+"),
+        StringConversionData("+--", L"+-"),
+
+        // the following are invalid UTF-7 sequences
+        StringConversionData("\xa3", NULL),
+        StringConversionData("+", NULL),
+        StringConversionData("+~", NULL),
+        StringConversionData("a+", NULL),
+    };
+
+    for ( size_t n = 0; n < WXSIZEOF(utf7data); n++ )
     {
     {
-        s1++;
-        s2++;
+        const StringConversionData& d = utf7data[n];
+
+        // converting to/from UTF-7 using iconv() currently doesn't work
+        // because of several problems:
+        //  - GetMBNulLen() doesn't return correct result (iconv converts L'\0'
+        //    to an incomplete and anyhow nonsensical "+AA" string)
+        //  - iconv refuses to convert "+-" (although it converts "+-\n" just
+        //    fine, go figure)
+        //
+        // I have no idea how to fix this so just disable the test for now
+#if 0
+        d.Test(n, wxCSConv("utf-7"));
+#endif
+        d.Test(n, wxConvUTF7);
     }
     }
-    return *s1 - *s2;
 }
 }
-#endif
 
 
-void
-UnicodeTestCase::DoTestConversion(const char *s,
-                                 const wchar_t *ws,
-                                 wxCSConv& conv)
+void UnicodeTestCase::ConversionUTF8()
 {
 {
-#if wxUSE_UNICODE
-    if ( ws )
+    static const StringConversionData utf8data[] =
     {
     {
-        wxCharBuffer buf(wxString(ws).mb_str(conv));
+#ifdef wxHAVE_U_ESCAPE
+        StringConversionData("\xc2\xa3", L"\u00a3"),
+#endif
+        StringConversionData("\xc2", NULL),
+    };
 
 
-        CPPUNIT_ASSERT( strcmp(buf, s) == 0 );
-    }
-#else // wxUSE_UNICODE
-    if ( s )
+    wxCSConv conv(wxT("utf-8"));
+    for ( size_t n = 0; n < WXSIZEOF(utf8data); n++ )
     {
     {
-        wxWCharBuffer wbuf(wxString(s).wc_str(conv));
-
-        if ( ws )
-            CPPUNIT_ASSERT( wx_wcscmp(wbuf, ws) == 0 );
-        else
-            CPPUNIT_ASSERT( !*wbuf );
+        const StringConversionData& d = utf8data[n];
+        d.Test(n, conv);
+        d.Test(n, wxConvUTF8);
     }
     }
-#endif // wxUSE_UNICODE/!wxUSE_UNICODE
 }
 
 }
 
-struct StringConversionData
-{
-    const char *str;
-    const wchar_t *wcs;
-};
-
-void UnicodeTestCase::ConversionUTF7()
+void UnicodeTestCase::ConversionUTF16()
 {
 {
-    static const StringConversionData utf7data[] =
+    static const StringConversionData utf16data[] =
     {
     {
-        { "+-", L"+" },
-        { "+--", L"+-" },
-        //\u isn't recognized on MSVC 6
-#if !defined(_MSC_VER)
-#if !defined(__GNUC__) || (__GNUC__ >= 3)
-        { "+AKM-", L"\u00a3" },
-#endif
+#ifdef wxHAVE_U_ESCAPE
+        StringConversionData(
+            "\x04\x1f\x04\x40\x04\x38\x04\x32\x04\x35\x04\x42\0\0",
+            L"\u041f\u0440\u0438\u0432\u0435\u0442"),
+        StringConversionData(
+            "\x01\0\0b\x01\0\0a\x01\0\0r\0\0",
+            L"\u0100b\u0100a\u0100r"),
 #endif
 #endif
-        // Windows accepts invalid UTF-7 strings and so does our UTF-7
-        // conversion code -- this is wrong IMO but the way it is for now
-        //
-        // notice that converting "+" still behaves as expected because the
-        // result is just an empty string, i.e. the same as if there were an
-        // error, but converting "a+" results in "a" while it really should
-        // fail
-        { "+", NULL },
-        { "a+", L"a" },
+        StringConversionData("\0f\0o\0o\0\0", L"foo"),
     };
 
     };
 
-    wxCSConv conv(_T("utf-7"));
-    for ( size_t n = 0; n < WXSIZEOF(utf7data); n++ )
+    wxCSConv conv(wxFONTENCODING_UTF16BE);
+    for ( size_t n = 0; n < WXSIZEOF(utf16data); n++ )
     {
     {
-        const StringConversionData& d = utf7data[n];
-        DoTestConversion(d.str, d.wcs, conv);
+        const StringConversionData& d = utf16data[n];
+        d.Test(n, conv);
     }
     }
+
+    // special case: this string has consecutive NULs inside it which don't
+    // terminate the string, this exposed a bug in our conversion code which
+    // got confused in this case
+    size_t len;
+    conv.cMB2WC("\x01\0\0B\0C" /* A macron BC */, 6, &len);
+    CPPUNIT_ASSERT_EQUAL( 3, len );
 }
 
 }
 
-void UnicodeTestCase::ConversionUTF8()
+void UnicodeTestCase::ConversionUTF32()
 {
 {
-    static const StringConversionData utf8data[] =
+    static const StringConversionData utf32data[] =
     {
     {
-        //\u isn't recognized on MSVC 6
-#if !defined(_MSC_VER)
-#if !defined(__GNUC__) || (__GNUC__ >= 3)
-        { "\xc2\xa3", L"\u00a3" },
-#endif
+#ifdef wxHAVE_U_ESCAPE
+        StringConversionData(
+            "\0\0\x04\x1f\0\0\x04\x40\0\0\x04\x38\0\0\x04\x32\0\0\x04\x35\0\0\x04\x42\0\0\0\0",
+          L"\u041f\u0440\u0438\u0432\u0435\u0442"),
 #endif
 #endif
-        { "\xc2", NULL },
+        StringConversionData("\0\0\0f\0\0\0o\0\0\0o\0\0\0\0", L"foo"),
     };
 
     };
 
-    wxCSConv conv(_T("utf-8"));
-    for ( size_t n = 0; n < WXSIZEOF(utf8data); n++ )
+    wxCSConv conv(wxFONTENCODING_UTF32BE);
+    for ( size_t n = 0; n < WXSIZEOF(utf32data); n++ )
     {
     {
-        const StringConversionData& d = utf8data[n];
-        DoTestConversion(d.str, d.wcs, conv);
+        const StringConversionData& d = utf32data[n];
+        d.Test(n, conv);
     }
     }
+
+    size_t len;
+    conv.cMB2WC("\0\0\x01\0\0\0\0B\0\0\0C" /* A macron BC */, 12, &len);
+    CPPUNIT_ASSERT_EQUAL( 3, len );
 }
 
 }
 
-#endif // wxUSE_WCHAR_T
+void UnicodeTestCase::IsConvOk()
+{
+    CPPUNIT_ASSERT( wxCSConv(wxFONTENCODING_SYSTEM).IsOk() );
+    CPPUNIT_ASSERT( wxCSConv("US-ASCII").IsOk() );
+    CPPUNIT_ASSERT( wxCSConv("UTF-8").IsOk() );
+    CPPUNIT_ASSERT( !wxCSConv("NoSuchConversion").IsOk() );
+
+#ifdef __WINDOWS__
+    CPPUNIT_ASSERT( wxCSConv("WINDOWS-437").IsOk() );
+#endif
+}
+
+#if wxUSE_UNICODE
+void UnicodeTestCase::Iteration()
+{
+    // "czech" in Czech ("cestina"):
+    static const char *textUTF8 = "\304\215e\305\241tina";
+    static const wchar_t textUTF16[] = {0x10D, 0x65, 0x161, 0x74, 0x69, 0x6E, 0x61, 0};
+
+    wxString text(wxString::FromUTF8(textUTF8));
+    CPPUNIT_ASSERT( wxStrcmp(text.wc_str(), textUTF16) == 0 );
+
+    // verify the string was decoded correctly:
+    {
+        size_t idx = 0;
+        for ( wxString::const_iterator i = text.begin(); i != text.end(); ++i, ++idx )
+        {
+            CPPUNIT_ASSERT( *i == textUTF16[idx] );
+        }
+    }
+
+    // overwrite the string with something that is shorter in UTF-8:
+    {
+        for ( wxString::iterator i = text.begin(); i != text.end(); ++i )
+            *i = 'x';
+    }
+
+    // restore the original text now:
+    {
+        wxString::iterator end1 = text.end();
+        wxString::const_iterator end2 = text.end();
+
+        size_t idx = 0;
+        for ( wxString::iterator i = text.begin(); i != text.end(); ++i, ++idx )
+        {
+            *i = textUTF16[idx];
+
+            CPPUNIT_ASSERT( end1 == text.end() );
+            CPPUNIT_ASSERT( end2 == text.end() );
+        }
+
+        CPPUNIT_ASSERT( end1 == text.end() );
+        CPPUNIT_ASSERT( end2 == text.end() );
+    }
+
+    // and verify it again:
+    {
+        size_t idx = 0;
+        for ( wxString::const_iterator i = text.begin(); i != text.end(); ++i, ++idx )
+        {
+            CPPUNIT_ASSERT( *i == textUTF16[idx] );
+        }
+    }
+}
+#endif // wxUSE_UNICODE