tests/strings/unicode.cpp

   1 ///////////////////////////////////////////////////////////////////////////////
   2 // Name:        tests/strings/unicode.cpp
   3 // Purpose:     Unicode unit test
   4 // Author:      Vadim Zeitlin, Wlodzimierz ABX Skiba
   5 // Created:     2004-04-28
   6 // RCS-ID:      $Id$
   7 // Copyright:   (c) 2004 Vadim Zeitlin, Wlodzimierz Skiba
   8 ///////////////////////////////////////////////////////////////////////////////
   9
  10 // ----------------------------------------------------------------------------
  11 // headers
  12 // ----------------------------------------------------------------------------
  13
  14 #include "testprec.h"
  15
  16 #ifdef __BORLANDC__
  17     #pragma hdrstop
  18 #endif
  19
  20 #ifndef WX_PRECOMP
  21     #include "wx/wx.h"
  22 #endif // WX_PRECOMP
  23
  24 // helper class holding the matching MB and WC strings
  25 //
  26 // either str or wcs (but not both) may be NULL, this means that the conversion
  27 // to it should fail
  28 struct StringConversionData
  29 {
  30     StringConversionData(const char *str_, const wchar_t *wcs_, int flags_ = 0)
  31         : str(str_), wcs(wcs_), flags(flags_)
  32     {
  33     }
  34
  35     const char * const str;
  36     const wchar_t * const wcs;
  37
  38     enum
  39     {
  40         TEST_BOTH  = 0, // test both str -> wcs and wcs -> str
  41         ONLY_MB2WC = 1  // only test str -> wcs conversion
  42     };
  43
  44     const int flags;
  45
  46     // test that the conversion between str and wcs (subject to flags) succeeds
  47     //
  48     // the first argument is the index in the test array and is used solely for
  49     // diagnostics
  50     void Test(size_t n, wxMBConv& conv) const
  51     {
  52         if ( str )
  53         {
  54             wxWCharBuffer wbuf = conv.cMB2WC(str);
  55
  56             if ( wcs )
  57             {
  58                 CPPUNIT_ASSERT_MESSAGE
  59                 (
  60                     Message(n, "MB2WC failed"),
  61                     wbuf.data()
  62                 );
  63
  64                 CPPUNIT_ASSERT_MESSAGE
  65                 (
  66                     Message(n, "MB2WC", wbuf, wcs),
  67                     wxStrcmp(wbuf, wcs) == 0
  68                 );
  69             }
  70             else // conversion is supposed to fail
  71             {
  72                 CPPUNIT_ASSERT_MESSAGE
  73                 (
  74                     Message(n, "MB2WC succeeded"),
  75                     !wbuf.data()
  76                 );
  77             }
  78         }
  79
  80         if ( wcs && !(flags & ONLY_MB2WC) )
  81         {
  82             wxCharBuffer buf = conv.cWC2MB(wcs);
  83
  84             if ( str )
  85             {
  86                 CPPUNIT_ASSERT_MESSAGE
  87                 (
  88                     Message(n, "WC2MB failed"),
  89                     buf.data()
  90                 );
  91
  92                 CPPUNIT_ASSERT_MESSAGE
  93                 (
  94                     Message(n, "WC2MB", buf, str),
  95                     strcmp(buf, str) == 0
  96                 );
  97             }
  98             else
  99             {
 100                 CPPUNIT_ASSERT_MESSAGE
 101                 (
 102                     Message(n, "WC2MB succeeded"),
 103                     !buf.data()
 104                 );
 105             }
 106         }
 107     }
 108
 109 private:
 110     static std::string
 111     Message(size_t n, const wxString& msg)
 112     {
 113         return std::string(wxString::Format("#%lu: %s", (unsigned long)n, msg));
 114     }
 115
 116     template <typename T>
 117     static std::string
 118     Message(size_t n,
 119             const char *func,
 120             const wxCharTypeBuffer<T>& actual,
 121             const T *expected)
 122     {
 123         return Message(n,
 124                        wxString::Format("%s returned \"%s\", expected \"%s\"",
 125                                         func, actual.data(), expected));
 126     }
 127 };
 128
 129 // ----------------------------------------------------------------------------
 130 // test class
 131 // ----------------------------------------------------------------------------
 132
 133 class UnicodeTestCase : public CppUnit::TestCase
 134 {
 135 public:
 136     UnicodeTestCase();
 137
 138 private:
 139     CPPUNIT_TEST_SUITE( UnicodeTestCase );
 140         CPPUNIT_TEST( ToFromAscii );
 141         CPPUNIT_TEST( ConstructorsWithConversion );
 142         CPPUNIT_TEST( ConversionEmpty );
 143         CPPUNIT_TEST( ConversionWithNULs );
 144         CPPUNIT_TEST( ConversionUTF7 );
 145         CPPUNIT_TEST( ConversionUTF8 );
 146         CPPUNIT_TEST( ConversionUTF16 );
 147         CPPUNIT_TEST( ConversionUTF32 );
 148         CPPUNIT_TEST( IsConvOk );
 149 #if wxUSE_UNICODE
 150         CPPUNIT_TEST( Iteration );
 151 #endif
 152     CPPUNIT_TEST_SUITE_END();
 153
 154     void ToFromAscii();
 155     void ConstructorsWithConversion();
 156     void ConversionEmpty();
 157     void ConversionWithNULs();
 158     void ConversionUTF7();
 159     void ConversionUTF8();
 160     void ConversionUTF16();
 161     void ConversionUTF32();
 162     void IsConvOk();
 163 #if wxUSE_UNICODE
 164     void Iteration();
 165 #endif
 166
 167     DECLARE_NO_COPY_CLASS(UnicodeTestCase)
 168 };
 169
 170 // register in the unnamed registry so that these tests are run by default
 171 CPPUNIT_TEST_SUITE_REGISTRATION( UnicodeTestCase );
 172
 173 // also include in it's own registry so that these tests can be run alone
 174 CPPUNIT_TEST_SUITE_NAMED_REGISTRATION( UnicodeTestCase, "UnicodeTestCase" );
 175
 176 UnicodeTestCase::UnicodeTestCase()
 177 {
 178 }
 179
 180 void UnicodeTestCase::ToFromAscii()
 181 {
 182
 183 #define TEST_TO_FROM_ASCII(txt)                              \
 184     {                                                        \
 185         static const char *msg = txt;                        \
 186         wxString s = wxString::FromAscii(msg);               \
 187         CPPUNIT_ASSERT( strcmp( s.ToAscii() , msg ) == 0 );  \
 188     }
 189
 190     TEST_TO_FROM_ASCII( "Hello, world!" );
 191     TEST_TO_FROM_ASCII( "additional \" special \t test \\ component \n :-)" );
 192 }
 193
 194 void UnicodeTestCase::ConstructorsWithConversion()
 195 {
 196     // the string "Déjà" in UTF-8 and wchar_t:
 197     const unsigned char utf8Buf[] = {0x44,0xC3,0xA9,0x6A,0xC3,0xA0,0};
 198     const unsigned char utf8subBuf[] = {0x44,0xC3,0xA9,0x6A,0}; // just "Déj"
 199     const char *utf8 = (char *)utf8Buf;
 200     const char *utf8sub = (char *)utf8subBuf;
 201
 202     wxString s1(utf8, wxConvUTF8);
 203
 204 #if wxUSE_UNICODE
 205     const wchar_t wchar[] = {0x44,0xE9,0x6A,0xE0,0};
 206     CPPUNIT_ASSERT_EQUAL( wchar, s1 );
 207
 208     wxString s2(wchar);
 209     CPPUNIT_ASSERT_EQUAL( wchar, s2 );
 210     CPPUNIT_ASSERT_EQUAL( wxString::FromUTF8(utf8), s2 );
 211 #else
 212     CPPUNIT_ASSERT_EQUAL( utf8, s1 );
 213 #endif
 214
 215     wxString sub(utf8sub, wxConvUTF8); // "Dej" substring
 216     wxString s3(utf8, wxConvUTF8, 4);
 217     CPPUNIT_ASSERT_EQUAL( sub, s3 );
 218
 219 #if wxUSE_UNICODE
 220     wxString s4(wchar, wxConvUTF8, 3);
 221     CPPUNIT_ASSERT_EQUAL( sub, s4 );
 222
 223     // conversion should stop with failure at pos 35
 224     wxString s("\t[pl]open.format.Sformatuj dyskietkê=gfloppy %f", wxConvUTF8);
 225     CPPUNIT_ASSERT( s.empty() );
 226 #endif // wxUSE_UNICODE
 227
 228
 229     // test using Unicode strings together with char* strings (this must work
 230     // in ANSI mode as well, of course):
 231     wxString s5("ascii");
 232     CPPUNIT_ASSERT_EQUAL( "ascii", s5 );
 233
 234     s5 += " value";
 235
 236     CPPUNIT_ASSERT( strcmp(s5.mb_str(), "ascii value") == 0 );
 237     CPPUNIT_ASSERT_EQUAL( "ascii value", s5 );
 238     CPPUNIT_ASSERT( s5 != "SomethingElse" );
 239 }
 240
 241 void UnicodeTestCase::ConversionEmpty()
 242 {
 243     size_t len;
 244
 245 #if wxUSE_UNICODE
 246     wxConvLibc.cWC2MB(L"", 0, &len);
 247 #else // !wxUSE_UNICODE
 248     wxConvLibc.cMB2WC("", 0, &len);
 249 #endif // wxUSE_UNICODE/!wxUSE_UNICODE
 250
 251     CPPUNIT_ASSERT_EQUAL( 0, len );
 252 }
 253
 254 void UnicodeTestCase::ConversionWithNULs()
 255 {
 256 #if wxUSE_UNICODE
 257     static const size_t lenNulString = 10;
 258
 259     wxString szTheString(L"The\0String", wxConvLibc, lenNulString);
 260     wxCharBuffer theBuffer = szTheString.mb_str();
 261
 262     CPPUNIT_ASSERT( memcmp(theBuffer.data(), "The\0String",
 263                     lenNulString + 1) == 0 );
 264
 265     wxString szTheString2("The\0String", wxConvLocal, lenNulString);
 266     CPPUNIT_ASSERT_EQUAL( lenNulString, szTheString2.length() );
 267     CPPUNIT_ASSERT( wxTmemcmp(szTheString2.c_str(), L"The\0String",
 268                     lenNulString + 1) == 0 );
 269 #else // !wxUSE_UNICODE
 270     wxString szTheString("TheString");
 271     szTheString.insert(3, 1, '\0');
 272     wxWCharBuffer theBuffer = szTheString.wc_str(wxConvLibc);
 273
 274     CPPUNIT_ASSERT( memcmp(theBuffer.data(), L"The\0String", 11 * sizeof(wchar_t)) == 0 );
 275
 276     wxString szLocalTheString("TheString");
 277     szLocalTheString.insert(3, 1, '\0');
 278     wxWCharBuffer theLocalBuffer = szLocalTheString.wc_str(wxConvLocal);
 279
 280     CPPUNIT_ASSERT( memcmp(theLocalBuffer.data(), L"The\0String", 11 * sizeof(wchar_t)) == 0 );
 281 #endif // wxUSE_UNICODE/!wxUSE_UNICODE
 282 }
 283
 284 void UnicodeTestCase::ConversionUTF7()
 285 {
 286     static const StringConversionData utf7data[] =
 287     {
 288         // normal fragments
 289         StringConversionData("+AKM-", L"\xa3"),
 290         StringConversionData("+AOk-t+AOk-", L"\xe9t\xe9"),
 291
 292         // this one is an alternative valid encoding of the same string
 293         StringConversionData("+AOk-t+AOk", L"\xe9t\xe9",
 294                              StringConversionData::ONLY_MB2WC),
 295
 296         // some special cases
 297         StringConversionData("+-", L"+"),
 298         StringConversionData("+--", L"+-"),
 299
 300         // the following are invalid UTF-7 sequences
 301         StringConversionData("\xa3", NULL),
 302         StringConversionData("+", NULL),
 303         StringConversionData("+~", NULL),
 304         StringConversionData("a+", NULL),
 305     };
 306
 307     for ( size_t n = 0; n < WXSIZEOF(utf7data); n++ )
 308     {
 309         const StringConversionData& d = utf7data[n];
 310
 311         // converting to/from UTF-7 using iconv() currently doesn't work
 312         // because of several problems:
 313         //  - GetMBNulLen() doesn't return correct result (iconv converts L'\0'
 314         //    to an incomplete and anyhow nonsensical "+AA" string)
 315         //  - iconv refuses to convert "+-" (although it converts "+-\n" just
 316         //    fine, go figure)
 317         //
 318         // I have no idea how to fix this so just disable the test for now
 319 #if 0
 320         d.Test(n, wxCSConv("utf-7"));
 321 #endif
 322         d.Test(n, wxConvUTF7);
 323     }
 324 }
 325
 326 void UnicodeTestCase::ConversionUTF8()
 327 {
 328     static const StringConversionData utf8data[] =
 329     {
 330 #ifdef wxHAVE_U_ESCAPE
 331         StringConversionData("\xc2\xa3", L"\u00a3"),
 332 #endif
 333         StringConversionData("\xc2", NULL),
 334     };
 335
 336     wxCSConv conv(_T("utf-8"));
 337     for ( size_t n = 0; n < WXSIZEOF(utf8data); n++ )
 338     {
 339         const StringConversionData& d = utf8data[n];
 340         d.Test(n, conv);
 341         d.Test(n, wxConvUTF8);
 342     }
 343 }
 344
 345 void UnicodeTestCase::ConversionUTF16()
 346 {
 347     static const StringConversionData utf16data[] =
 348     {
 349 #ifdef wxHAVE_U_ESCAPE
 350         StringConversionData(
 351             "\x04\x1f\x04\x40\x04\x38\x04\x32\x04\x35\x04\x42\0\0",
 352             L"\u041f\u0440\u0438\u0432\u0435\u0442"),
 353         StringConversionData(
 354             "\x01\0\0b\x01\0\0a\x01\0\0r\0\0",
 355             L"\u0100b\u0100a\u0100r"),
 356 #endif
 357         StringConversionData("\0f\0o\0o\0\0", L"foo"),
 358     };
 359
 360     wxCSConv conv(wxFONTENCODING_UTF16BE);
 361     for ( size_t n = 0; n < WXSIZEOF(utf16data); n++ )
 362     {
 363         const StringConversionData& d = utf16data[n];
 364         d.Test(n, conv);
 365     }
 366
 367     // special case: this string has consecutive NULs inside it which don't
 368     // terminate the string, this exposed a bug in our conversion code which
 369     // got confused in this case
 370     size_t len;
 371     conv.cMB2WC("\x01\0\0B\0C" /* A macron BC */, 6, &len);
 372     CPPUNIT_ASSERT_EQUAL( 3, len );
 373 }
 374
 375 void UnicodeTestCase::ConversionUTF32()
 376 {
 377     static const StringConversionData utf32data[] =
 378     {
 379 #ifdef wxHAVE_U_ESCAPE
 380         StringConversionData(
 381             "\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",
 382           L"\u041f\u0440\u0438\u0432\u0435\u0442"),
 383 #endif
 384         StringConversionData("\0\0\0f\0\0\0o\0\0\0o\0\0\0\0", L"foo"),
 385     };
 386
 387     wxCSConv conv(wxFONTENCODING_UTF32BE);
 388     for ( size_t n = 0; n < WXSIZEOF(utf32data); n++ )
 389     {
 390         const StringConversionData& d = utf32data[n];
 391         d.Test(n, conv);
 392     }
 393
 394     size_t len;
 395     conv.cMB2WC("\0\0\x01\0\0\0\0B\0\0\0C" /* A macron BC */, 12, &len);
 396     CPPUNIT_ASSERT_EQUAL( 3, len );
 397 }
 398
 399 void UnicodeTestCase::IsConvOk()
 400 {
 401     CPPUNIT_ASSERT( wxCSConv(wxFONTENCODING_SYSTEM).IsOk() );
 402     CPPUNIT_ASSERT( wxCSConv("US-ASCII").IsOk() );
 403     CPPUNIT_ASSERT( wxCSConv("UTF-8").IsOk() );
 404     CPPUNIT_ASSERT( !wxCSConv("NoSuchConversion").IsOk() );
 405
 406 #ifdef __WINDOWS__
 407     CPPUNIT_ASSERT( wxCSConv("WINDOWS-437").IsOk() );
 408 #endif
 409 }
 410
 411 #if wxUSE_UNICODE
 412 void UnicodeTestCase::Iteration()
 413 {
 414     // "czech" in Czech ("cestina"):
 415     static const char *textUTF8 = "\304\215e\305\241tina";
 416     static const wchar_t textUTF16[] = {0x10D, 0x65, 0x161, 0x74, 0x69, 0x6E, 0x61, 0};
 417
 418     wxString text(wxString::FromUTF8(textUTF8));
 419     CPPUNIT_ASSERT( wxStrcmp(text.wc_str(), textUTF16) == 0 );
 420
 421     // verify the string was decoded correctly:
 422     {
 423         size_t idx = 0;
 424         for ( wxString::const_iterator i = text.begin(); i != text.end(); ++i, ++idx )
 425         {
 426             CPPUNIT_ASSERT( *i == textUTF16[idx] );
 427         }
 428     }
 429
 430     // overwrite the string with something that is shorter in UTF-8:
 431     {
 432         for ( wxString::iterator i = text.begin(); i != text.end(); ++i )
 433             *i = 'x';
 434     }
 435
 436     // restore the original text now:
 437     {
 438         wxString::iterator end1 = text.end();
 439         wxString::const_iterator end2 = text.end();
 440
 441         size_t idx = 0;
 442         for ( wxString::iterator i = text.begin(); i != text.end(); ++i, ++idx )
 443         {
 444             *i = textUTF16[idx];
 445
 446             CPPUNIT_ASSERT( end1 == text.end() );
 447             CPPUNIT_ASSERT( end2 == text.end() );
 448         }
 449
 450         CPPUNIT_ASSERT( end1 == text.end() );
 451         CPPUNIT_ASSERT( end2 == text.end() );
 452     }
 453
 454     // and verify it again:
 455     {
 456         size_t idx = 0;
 457         for ( wxString::const_iterator i = text.begin(); i != text.end(); ++i, ++idx )
 458         {
 459             CPPUNIT_ASSERT( *i == textUTF16[idx] );
 460         }
 461     }
 462 }
 463 #endif // wxUSE_UNICODE