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