tests/benchmarks/htmlparser/htmlpars.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        src/html/htmlpars.cpp
   3 // Purpose:     wx28HtmlParser class (generic parser)
   4 // Author:      Vaclav Slavik
   5 // Copyright:   (c) 1999 Vaclav Slavik
   6 // Licence:     wxWindows licence
   7 /////////////////////////////////////////////////////////////////////////////
   8
   9 #include "wx/wxprec.h"
  10
  11 #ifdef __BORLANDC__
  12     #pragma hdrstop
  13 #endif
  14
  15 #include "htmlpars.h"
  16
  17 #ifndef WXPRECOMP
  18     #include "wx/dynarray.h"
  19     #include "wx/log.h"
  20     #include "wx/intl.h"
  21     #include "wx/app.h"
  22 #endif
  23
  24 #include "wx/tokenzr.h"
  25 #include "wx/wfstream.h"
  26 #include "wx/url.h"
  27 #include "wx/fontmap.h"
  28 #include "wx/html/htmldefs.h"
  29 #include "wx/arrimpl.cpp"
  30
  31 #ifdef __WXWINCE__
  32     #include "wx/msw/wince/missing.h"       // for bsearch()
  33 #endif
  34
  35 // DLL options compatibility check:
  36 WX_CHECK_BUILD_OPTIONS("wxHTML")
  37
  38 const wxChar *wxTRACE_HTML_DEBUG = wxT("htmldebug");
  39
  40 //-----------------------------------------------------------------------------
  41 // wx28HtmlParser helpers
  42 //-----------------------------------------------------------------------------
  43
  44 class wx28HtmlTextPiece
  45 {
  46 public:
  47     wx28HtmlTextPiece(int pos, int lng) : m_pos(pos), m_lng(lng) {}
  48     int m_pos, m_lng;
  49 };
  50
  51 WX_DECLARE_OBJARRAY(wx28HtmlTextPiece, wx28HtmlTextPieces);
  52 WX_DEFINE_OBJARRAY(wx28HtmlTextPieces)
  53
  54 class wx28HtmlParserState
  55 {
  56 public:
  57     wx28HtmlTag         *m_curTag;
  58     wx28HtmlTag         *m_tags;
  59     wx28HtmlTextPieces  *m_textPieces;
  60     int                m_curTextPiece;
  61     wxString           m_source;
  62     wx28HtmlParserState *m_nextState;
  63 };
  64
  65 //-----------------------------------------------------------------------------
  66 // wx28HtmlParser
  67 //-----------------------------------------------------------------------------
  68
  69 IMPLEMENT_ABSTRACT_CLASS(wx28HtmlParser,wxObject)
  70
  71 wx28HtmlParser::wx28HtmlParser()
  72     : wxObject(), m_HandlersHash(wxKEY_STRING),
  73       m_FS(NULL), m_HandlersStack(NULL)
  74 {
  75     m_entitiesParser = new wx28HtmlEntitiesParser;
  76     m_Tags = NULL;
  77     m_CurTag = NULL;
  78     m_TextPieces = NULL;
  79     m_CurTextPiece = 0;
  80     m_SavedStates = NULL;
  81 }
  82
  83 wx28HtmlParser::~wx28HtmlParser()
  84 {
  85     while (RestoreState()) {}
  86     DestroyDOMTree();
  87
  88     if (m_HandlersStack)
  89     {
  90         wxList& tmp = *m_HandlersStack;
  91         wxList::iterator it, en;
  92         for( it = tmp.begin(), en = tmp.end(); it != en; ++it )
  93             delete (wxHashTable*)*it;
  94         tmp.clear();
  95     }
  96     delete m_HandlersStack;
  97     m_HandlersHash.Clear();
  98     WX_CLEAR_LIST(wxList, m_HandlersList);
  99     delete m_entitiesParser;
 100 }
 101
 102 wxObject* wx28HtmlParser::Parse(const wxString& source)
 103 {
 104     InitParser(source);
 105     DoParsing();
 106     wxObject *result = GetProduct();
 107     DoneParser();
 108     return result;
 109 }
 110
 111 void wx28HtmlParser::InitParser(const wxString& source)
 112 {
 113     SetSource(source);
 114     m_stopParsing = false;
 115 }
 116
 117 void wx28HtmlParser::DoneParser()
 118 {
 119     DestroyDOMTree();
 120 }
 121
 122 void wx28HtmlParser::SetSource(const wxString& src)
 123 {
 124     DestroyDOMTree();
 125     m_Source = src;
 126     CreateDOMTree();
 127     m_CurTag = NULL;
 128     m_CurTextPiece = 0;
 129 }
 130
 131 void wx28HtmlParser::CreateDOMTree()
 132 {
 133     wx28HtmlTagsCache cache(m_Source);
 134     m_TextPieces = new wx28HtmlTextPieces;
 135     CreateDOMSubTree(NULL, 0, m_Source.length(), &cache);
 136     m_CurTextPiece = 0;
 137 }
 138
 139 extern bool wxIsCDATAElement(const wxChar *tag);
 140
 141 void wx28HtmlParser::CreateDOMSubTree(wx28HtmlTag *cur,
 142                                     int begin_pos, int end_pos,
 143                                     wx28HtmlTagsCache *cache)
 144 {
 145     if (end_pos <= begin_pos) return;
 146
 147     wxChar c;
 148     int i = begin_pos;
 149     int textBeginning = begin_pos;
 150
 151     // If the tag contains CDATA text, we include the text between beginning
 152     // and ending tag verbosely. Setting i=end_pos will skip to the very
 153     // end of this function where text piece is added, bypassing any child
 154     // tags parsing (CDATA element can't have child elements by definition):
 155     if (cur != NULL && wxIsCDATAElement(cur->GetName().c_str()))
 156     {
 157         i = end_pos;
 158     }
 159
 160     while (i < end_pos)
 161     {
 162         c = m_Source.GetChar(i);
 163
 164         if (c == wxT('<'))
 165         {
 166             // add text to m_TextPieces:
 167             if (i - textBeginning > 0)
 168                 m_TextPieces->Add(
 169                     wx28HtmlTextPiece(textBeginning, i - textBeginning));
 170
 171             // if it is a comment, skip it:
 172             if (i < end_pos-6 && m_Source.GetChar(i+1) == wxT('!') &&
 173                                  m_Source.GetChar(i+2) == wxT('-') &&
 174                                  m_Source.GetChar(i+3) == wxT('-'))
 175             {
 176                 // Comments begin with "<!--" and end with "--[ \t\r\n]*>"
 177                 // according to HTML 4.0
 178                 int dashes = 0;
 179                 i += 4;
 180                 while (i < end_pos)
 181                 {
 182                     c = m_Source.GetChar(i++);
 183                     if ((c == wxT(' ') || c == wxT('\n') ||
 184                         c == wxT('\r') || c == wxT('\t')) && dashes >= 2) {}
 185                     else if (c == wxT('>') && dashes >= 2)
 186                     {
 187                         textBeginning = i;
 188                         break;
 189                     }
 190                     else if (c == wxT('-'))
 191                         dashes++;
 192                     else
 193                         dashes = 0;
 194                 }
 195             }
 196
 197             // add another tag to the tree:
 198             else if (i < end_pos-1 && m_Source.GetChar(i+1) != wxT('/'))
 199             {
 200                 wx28HtmlTag *chd;
 201                 if (cur)
 202                     chd = new wx28HtmlTag(cur, m_Source,
 203                                         i, end_pos, cache, m_entitiesParser);
 204                 else
 205                 {
 206                     chd = new wx28HtmlTag(NULL, m_Source,
 207                                         i, end_pos, cache, m_entitiesParser);
 208                     if (!m_Tags)
 209                     {
 210                         // if this is the first tag to be created make the root
 211                         // m_Tags point to it:
 212                         m_Tags = chd;
 213                     }
 214                     else
 215                     {
 216                         // if there is already a root tag add this tag as
 217                         // the last sibling:
 218                         chd->m_Prev = m_Tags->GetLastSibling();
 219                         chd->m_Prev->m_Next = chd;
 220                     }
 221                 }
 222
 223                 if (chd->HasEnding())
 224                 {
 225                     CreateDOMSubTree(chd,
 226                                      chd->GetBeginPos(), chd->GetEndPos1(),
 227                                      cache);
 228                     i = chd->GetEndPos2();
 229                 }
 230                 else
 231                     i = chd->GetBeginPos();
 232
 233                 textBeginning = i;
 234             }
 235
 236             // ... or skip ending tag:
 237             else
 238             {
 239                 while (i < end_pos && m_Source.GetChar(i) != wxT('>')) i++;
 240                 textBeginning = i+1;
 241             }
 242         }
 243         else i++;
 244     }
 245
 246     // add remaining text to m_TextPieces:
 247     if (end_pos - textBeginning > 0)
 248         m_TextPieces->Add(
 249             wx28HtmlTextPiece(textBeginning, end_pos - textBeginning));
 250 }
 251
 252 void wx28HtmlParser::DestroyDOMTree()
 253 {
 254     wx28HtmlTag *t1, *t2;
 255     t1 = m_Tags;
 256     while (t1)
 257     {
 258         t2 = t1->GetNextSibling();
 259         delete t1;
 260         t1 = t2;
 261     }
 262     m_Tags = m_CurTag = NULL;
 263
 264     delete m_TextPieces;
 265     m_TextPieces = NULL;
 266 }
 267
 268 void wx28HtmlParser::DoParsing()
 269 {
 270     m_CurTag = m_Tags;
 271     m_CurTextPiece = 0;
 272     DoParsing(0, m_Source.length());
 273 }
 274
 275 void wx28HtmlParser::DoParsing(int begin_pos, int end_pos)
 276 {
 277     if (end_pos <= begin_pos) return;
 278
 279     wx28HtmlTextPieces& pieces = *m_TextPieces;
 280     size_t piecesCnt = pieces.GetCount();
 281
 282     while (begin_pos < end_pos)
 283     {
 284         while (m_CurTag && m_CurTag->GetBeginPos() < begin_pos)
 285             m_CurTag = m_CurTag->GetNextTag();
 286         while (m_CurTextPiece < piecesCnt &&
 287                pieces[m_CurTextPiece].m_pos < begin_pos)
 288             m_CurTextPiece++;
 289
 290         if (m_CurTextPiece < piecesCnt &&
 291             (!m_CurTag ||
 292              pieces[m_CurTextPiece].m_pos < m_CurTag->GetBeginPos()))
 293         {
 294             // Add text:
 295             AddText(GetEntitiesParser()->Parse(
 296                        m_Source.Mid(pieces[m_CurTextPiece].m_pos,
 297                                     pieces[m_CurTextPiece].m_lng)).t_str());
 298             begin_pos = pieces[m_CurTextPiece].m_pos +
 299                         pieces[m_CurTextPiece].m_lng;
 300             m_CurTextPiece++;
 301         }
 302         else if (m_CurTag)
 303         {
 304             if (m_CurTag->HasEnding())
 305                 begin_pos = m_CurTag->GetEndPos2();
 306             else
 307                 begin_pos = m_CurTag->GetBeginPos();
 308             wx28HtmlTag *t = m_CurTag;
 309             m_CurTag = m_CurTag->GetNextTag();
 310             AddTag(*t);
 311             if (m_stopParsing)
 312                 return;
 313         }
 314         else break;
 315     }
 316 }
 317
 318 void wx28HtmlParser::AddTag(const wx28HtmlTag& tag)
 319 {
 320     wx28HtmlTagHandler *h;
 321     bool inner = false;
 322
 323     h = (wx28HtmlTagHandler*) m_HandlersHash.Get(tag.GetName());
 324     if (h)
 325     {
 326         inner = h->HandleTag(tag);
 327         if (m_stopParsing)
 328             return;
 329     }
 330     if (!inner)
 331     {
 332         if (tag.HasEnding())
 333             DoParsing(tag.GetBeginPos(), tag.GetEndPos1());
 334     }
 335 }
 336
 337 void wx28HtmlParser::AddTagHandler(wx28HtmlTagHandler *handler)
 338 {
 339     wxString s(handler->GetSupportedTags());
 340     wxStringTokenizer tokenizer(s, wxT(", "));
 341
 342     while (tokenizer.HasMoreTokens())
 343         m_HandlersHash.Put(tokenizer.GetNextToken(), handler);
 344
 345     if (m_HandlersList.IndexOf(handler) == wxNOT_FOUND)
 346         m_HandlersList.Append(handler);
 347
 348     handler->SetParser(this);
 349 }
 350
 351 void wx28HtmlParser::PushTagHandler(wx28HtmlTagHandler *handler, const wxString& tags)
 352 {
 353     wxStringTokenizer tokenizer(tags, wxT(", "));
 354     wxString key;
 355
 356     if (m_HandlersStack == NULL)
 357     {
 358         m_HandlersStack = new wxList;
 359     }
 360
 361     m_HandlersStack->Insert((wxObject*)new wxHashTable(m_HandlersHash));
 362
 363     while (tokenizer.HasMoreTokens())
 364     {
 365         key = tokenizer.GetNextToken();
 366         m_HandlersHash.Delete(key);
 367         m_HandlersHash.Put(key, handler);
 368     }
 369 }
 370
 371 void wx28HtmlParser::PopTagHandler()
 372 {
 373     wxList::compatibility_iterator first;
 374
 375     if ( !m_HandlersStack ||
 376 #if wxUSE_STL
 377          !(first = m_HandlersStack->GetFirst())
 378 #else // !wxUSE_STL
 379          ((first = m_HandlersStack->GetFirst()) == NULL)
 380 #endif // wxUSE_STL/!wxUSE_STL
 381         )
 382     {
 383         wxLogWarning(_("Warning: attempt to remove HTML tag handler from empty stack."));
 384         return;
 385     }
 386     m_HandlersHash = *((wxHashTable*) first->GetData());
 387     delete (wxHashTable*) first->GetData();
 388     m_HandlersStack->Erase(first);
 389 }
 390
 391 void wx28HtmlParser::SetSourceAndSaveState(const wxString& src)
 392 {
 393     wx28HtmlParserState *s = new wx28HtmlParserState;
 394
 395     s->m_curTag = m_CurTag;
 396     s->m_tags = m_Tags;
 397     s->m_textPieces = m_TextPieces;
 398     s->m_curTextPiece = m_CurTextPiece;
 399     s->m_source = m_Source;
 400
 401     s->m_nextState = m_SavedStates;
 402     m_SavedStates = s;
 403
 404     m_CurTag = NULL;
 405     m_Tags = NULL;
 406     m_TextPieces = NULL;
 407     m_CurTextPiece = 0;
 408     m_Source = wxEmptyString;
 409
 410     SetSource(src);
 411 }
 412
 413 bool wx28HtmlParser::RestoreState()
 414 {
 415     if (!m_SavedStates) return false;
 416
 417     DestroyDOMTree();
 418
 419     wx28HtmlParserState *s = m_SavedStates;
 420     m_SavedStates = s->m_nextState;
 421
 422     m_CurTag = s->m_curTag;
 423     m_Tags = s->m_tags;
 424     m_TextPieces = s->m_textPieces;
 425     m_CurTextPiece = s->m_curTextPiece;
 426     m_Source = s->m_source;
 427
 428     delete s;
 429     return true;
 430 }
 431
 432 wxString wx28HtmlParser::GetInnerSource(const wx28HtmlTag& tag)
 433 {
 434     return GetSource()->Mid(tag.GetBeginPos(),
 435                             tag.GetEndPos1() - tag.GetBeginPos());
 436 }
 437
 438 //-----------------------------------------------------------------------------
 439 // wx28HtmlTagHandler
 440 //-----------------------------------------------------------------------------
 441
 442 IMPLEMENT_ABSTRACT_CLASS(wx28HtmlTagHandler,wxObject)
 443
 444 void wx28HtmlTagHandler::ParseInnerSource(const wxString& source)
 445 {
 446     // It is safe to temporarily change the source being parsed,
 447     // provided we restore the state back after parsing
 448     m_Parser->SetSourceAndSaveState(source);
 449     m_Parser->DoParsing();
 450     m_Parser->RestoreState();
 451 }
 452
 453
 454 //-----------------------------------------------------------------------------
 455 // wx28HtmlEntitiesParser
 456 //-----------------------------------------------------------------------------
 457
 458 IMPLEMENT_DYNAMIC_CLASS(wx28HtmlEntitiesParser,wxObject)
 459
 460 wx28HtmlEntitiesParser::wx28HtmlEntitiesParser()
 461 #if !wxUSE_UNICODE
 462     : m_conv(NULL), m_encoding(wxFONTENCODING_SYSTEM)
 463 #endif
 464 {
 465 }
 466
 467 wx28HtmlEntitiesParser::~wx28HtmlEntitiesParser()
 468 {
 469 #if !wxUSE_UNICODE
 470     delete m_conv;
 471 #endif
 472 }
 473
 474 void wx28HtmlEntitiesParser::SetEncoding(wxFontEncoding encoding)
 475 {
 476 #if !wxUSE_UNICODE
 477     if (encoding == m_encoding)
 478         return;
 479
 480     delete m_conv;
 481
 482     m_encoding = encoding;
 483     if (m_encoding == wxFONTENCODING_SYSTEM)
 484         m_conv = NULL;
 485     else
 486         m_conv = new wxCSConv(wxFontMapper::GetEncodingName(m_encoding));
 487 #else
 488     (void) encoding;
 489 #endif
 490 }
 491
 492 wxString wx28HtmlEntitiesParser::Parse(const wxString& input)
 493 {
 494     const wxChar *c, *last;
 495     const wxChar *in_str = input.c_str();
 496     wxString output;
 497
 498     for (c = in_str, last = in_str; *c != wxT('\0'); c++)
 499     {
 500         if (*c == wxT('&'))
 501         {
 502             if ( output.empty() )
 503                 output.reserve(input.length());
 504
 505             if (c - last > 0)
 506                 output.append(last, c - last);
 507             if ( *++c == wxT('\0') )
 508                 break;
 509
 510             wxString entity;
 511             const wxChar *ent_s = c;
 512             wxChar entity_char;
 513
 514             for (; (*c >= wxT('a') && *c <= wxT('z')) ||
 515                    (*c >= wxT('A') && *c <= wxT('Z')) ||
 516                    (*c >= wxT('0') && *c <= wxT('9')) ||
 517                    *c == wxT('_') || *c == wxT('#'); c++) {}
 518             entity.append(ent_s, c - ent_s);
 519             if (*c != wxT(';')) c--;
 520             last = c+1;
 521             entity_char = GetEntityChar(entity);
 522             if (entity_char)
 523                 output << entity_char;
 524             else
 525             {
 526                 output.append(ent_s-1, c-ent_s+2);
 527                 wxLogTrace(wxTRACE_HTML_DEBUG,
 528                            wxT("Unrecognized HTML entity: '%s'"),
 529                            entity.c_str());
 530             }
 531         }
 532     }
 533     if (last == in_str) // common case: no entity
 534         return input;
 535     if (*last != wxT('\0'))
 536         output.append(last);
 537     return output;
 538 }
 539
 540 struct wx28HtmlEntityInfo
 541 {
 542     const wxChar *name;
 543     unsigned code;
 544 };
 545
 546 extern "C" int LINKAGEMODE wx28HtmlEntityCompare(const void *key, const void *item)
 547 {
 548     return wxStrcmp((wxChar*)key, ((wx28HtmlEntityInfo*)item)->name);
 549 }
 550
 551 #if !wxUSE_UNICODE
 552 wxChar wx28HtmlEntitiesParser::GetCharForCode(unsigned code)
 553 {
 554     char buf[2];
 555     wchar_t wbuf[2];
 556     wbuf[0] = (wchar_t)code;
 557     wbuf[1] = 0;
 558     wxMBConv *conv = m_conv ? m_conv : &wxConvLocal;
 559     if (conv->WC2MB(buf, wbuf, 2) == (size_t)-1)
 560         return '?';
 561     return buf[0];
 562 }
 563 #endif
 564
 565 wxChar wx28HtmlEntitiesParser::GetEntityChar(const wxString& entity)
 566 {
 567     unsigned code = 0;
 568
 569     if (entity[0] == wxT('#'))
 570     {
 571         const wxChar *ent_s = entity.c_str();
 572         const wxChar *format;
 573
 574         if (ent_s[1] == wxT('x') || ent_s[1] == wxT('X'))
 575         {
 576             format = wxT("%x");
 577             ent_s++;
 578         }
 579         else
 580             format = wxT("%u");
 581         ent_s++;
 582
 583         if (wxSscanf(ent_s, format, &code) != 1)
 584             code = 0;
 585     }
 586     else
 587     {
 588         static wx28HtmlEntityInfo substitutions[] = {
 589             { wxT("AElig"),198 },
 590             { wxT("Aacute"),193 },
 591             { wxT("Acirc"),194 },
 592             { wxT("Agrave"),192 },
 593             { wxT("Alpha"),913 },
 594             { wxT("Aring"),197 },
 595             { wxT("Atilde"),195 },
 596             { wxT("Auml"),196 },
 597             { wxT("Beta"),914 },
 598             { wxT("Ccedil"),199 },
 599             { wxT("Chi"),935 },
 600             { wxT("Dagger"),8225 },
 601             { wxT("Delta"),916 },
 602             { wxT("ETH"),208 },
 603             { wxT("Eacute"),201 },
 604             { wxT("Ecirc"),202 },
 605             { wxT("Egrave"),200 },
 606             { wxT("Epsilon"),917 },
 607             { wxT("Eta"),919 },
 608             { wxT("Euml"),203 },
 609             { wxT("Gamma"),915 },
 610             { wxT("Iacute"),205 },
 611             { wxT("Icirc"),206 },
 612             { wxT("Igrave"),204 },
 613             { wxT("Iota"),921 },
 614             { wxT("Iuml"),207 },
 615             { wxT("Kappa"),922 },
 616             { wxT("Lambda"),923 },
 617             { wxT("Mu"),924 },
 618             { wxT("Ntilde"),209 },
 619             { wxT("Nu"),925 },
 620             { wxT("OElig"),338 },
 621             { wxT("Oacute"),211 },
 622             { wxT("Ocirc"),212 },
 623             { wxT("Ograve"),210 },
 624             { wxT("Omega"),937 },
 625             { wxT("Omicron"),927 },
 626             { wxT("Oslash"),216 },
 627             { wxT("Otilde"),213 },
 628             { wxT("Ouml"),214 },
 629             { wxT("Phi"),934 },
 630             { wxT("Pi"),928 },
 631             { wxT("Prime"),8243 },
 632             { wxT("Psi"),936 },
 633             { wxT("Rho"),929 },
 634             { wxT("Scaron"),352 },
 635             { wxT("Sigma"),931 },
 636             { wxT("THORN"),222 },
 637             { wxT("Tau"),932 },
 638             { wxT("Theta"),920 },
 639             { wxT("Uacute"),218 },
 640             { wxT("Ucirc"),219 },
 641             { wxT("Ugrave"),217 },
 642             { wxT("Upsilon"),933 },
 643             { wxT("Uuml"),220 },
 644             { wxT("Xi"),926 },
 645             { wxT("Yacute"),221 },
 646             { wxT("Yuml"),376 },
 647             { wxT("Zeta"),918 },
 648             { wxT("aacute"),225 },
 649             { wxT("acirc"),226 },
 650             { wxT("acute"),180 },
 651             { wxT("aelig"),230 },
 652             { wxT("agrave"),224 },
 653             { wxT("alefsym"),8501 },
 654             { wxT("alpha"),945 },
 655             { wxT("amp"),38 },
 656             { wxT("and"),8743 },
 657             { wxT("ang"),8736 },
 658             { wxT("aring"),229 },
 659             { wxT("asymp"),8776 },
 660             { wxT("atilde"),227 },
 661             { wxT("auml"),228 },
 662             { wxT("bdquo"),8222 },
 663             { wxT("beta"),946 },
 664             { wxT("brvbar"),166 },
 665             { wxT("bull"),8226 },
 666             { wxT("cap"),8745 },
 667             { wxT("ccedil"),231 },
 668             { wxT("cedil"),184 },
 669             { wxT("cent"),162 },
 670             { wxT("chi"),967 },
 671             { wxT("circ"),710 },
 672             { wxT("clubs"),9827 },
 673             { wxT("cong"),8773 },
 674             { wxT("copy"),169 },
 675             { wxT("crarr"),8629 },
 676             { wxT("cup"),8746 },
 677             { wxT("curren"),164 },
 678             { wxT("dArr"),8659 },
 679             { wxT("dagger"),8224 },
 680             { wxT("darr"),8595 },
 681             { wxT("deg"),176 },
 682             { wxT("delta"),948 },
 683             { wxT("diams"),9830 },
 684             { wxT("divide"),247 },
 685             { wxT("eacute"),233 },
 686             { wxT("ecirc"),234 },
 687             { wxT("egrave"),232 },
 688             { wxT("empty"),8709 },
 689             { wxT("emsp"),8195 },
 690             { wxT("ensp"),8194 },
 691             { wxT("epsilon"),949 },
 692             { wxT("equiv"),8801 },
 693             { wxT("eta"),951 },
 694             { wxT("eth"),240 },
 695             { wxT("euml"),235 },
 696             { wxT("euro"),8364 },
 697             { wxT("exist"),8707 },
 698             { wxT("fnof"),402 },
 699             { wxT("forall"),8704 },
 700             { wxT("frac12"),189 },
 701             { wxT("frac14"),188 },
 702             { wxT("frac34"),190 },
 703             { wxT("frasl"),8260 },
 704             { wxT("gamma"),947 },
 705             { wxT("ge"),8805 },
 706             { wxT("gt"),62 },
 707             { wxT("hArr"),8660 },
 708             { wxT("harr"),8596 },
 709             { wxT("hearts"),9829 },
 710             { wxT("hellip"),8230 },
 711             { wxT("iacute"),237 },
 712             { wxT("icirc"),238 },
 713             { wxT("iexcl"),161 },
 714             { wxT("igrave"),236 },
 715             { wxT("image"),8465 },
 716             { wxT("infin"),8734 },
 717             { wxT("int"),8747 },
 718             { wxT("iota"),953 },
 719             { wxT("iquest"),191 },
 720             { wxT("isin"),8712 },
 721             { wxT("iuml"),239 },
 722             { wxT("kappa"),954 },
 723             { wxT("lArr"),8656 },
 724             { wxT("lambda"),955 },
 725             { wxT("lang"),9001 },
 726             { wxT("laquo"),171 },
 727             { wxT("larr"),8592 },
 728             { wxT("lceil"),8968 },
 729             { wxT("ldquo"),8220 },
 730             { wxT("le"),8804 },
 731             { wxT("lfloor"),8970 },
 732             { wxT("lowast"),8727 },
 733             { wxT("loz"),9674 },
 734             { wxT("lrm"),8206 },
 735             { wxT("lsaquo"),8249 },
 736             { wxT("lsquo"),8216 },
 737             { wxT("lt"),60 },
 738             { wxT("macr"),175 },
 739             { wxT("mdash"),8212 },
 740             { wxT("micro"),181 },
 741             { wxT("middot"),183 },
 742             { wxT("minus"),8722 },
 743             { wxT("mu"),956 },
 744             { wxT("nabla"),8711 },
 745             { wxT("nbsp"),160 },
 746             { wxT("ndash"),8211 },
 747             { wxT("ne"),8800 },
 748             { wxT("ni"),8715 },
 749             { wxT("not"),172 },
 750             { wxT("notin"),8713 },
 751             { wxT("nsub"),8836 },
 752             { wxT("ntilde"),241 },
 753             { wxT("nu"),957 },
 754             { wxT("oacute"),243 },
 755             { wxT("ocirc"),244 },
 756             { wxT("oelig"),339 },
 757             { wxT("ograve"),242 },
 758             { wxT("oline"),8254 },
 759             { wxT("omega"),969 },
 760             { wxT("omicron"),959 },
 761             { wxT("oplus"),8853 },
 762             { wxT("or"),8744 },
 763             { wxT("ordf"),170 },
 764             { wxT("ordm"),186 },
 765             { wxT("oslash"),248 },
 766             { wxT("otilde"),245 },
 767             { wxT("otimes"),8855 },
 768             { wxT("ouml"),246 },
 769             { wxT("para"),182 },
 770             { wxT("part"),8706 },
 771             { wxT("permil"),8240 },
 772             { wxT("perp"),8869 },
 773             { wxT("phi"),966 },
 774             { wxT("pi"),960 },
 775             { wxT("piv"),982 },
 776             { wxT("plusmn"),177 },
 777             { wxT("pound"),163 },
 778             { wxT("prime"),8242 },
 779             { wxT("prod"),8719 },
 780             { wxT("prop"),8733 },
 781             { wxT("psi"),968 },
 782             { wxT("quot"),34 },
 783             { wxT("rArr"),8658 },
 784             { wxT("radic"),8730 },
 785             { wxT("rang"),9002 },
 786             { wxT("raquo"),187 },
 787             { wxT("rarr"),8594 },
 788             { wxT("rceil"),8969 },
 789             { wxT("rdquo"),8221 },
 790             { wxT("real"),8476 },
 791             { wxT("reg"),174 },
 792             { wxT("rfloor"),8971 },
 793             { wxT("rho"),961 },
 794             { wxT("rlm"),8207 },
 795             { wxT("rsaquo"),8250 },
 796             { wxT("rsquo"),8217 },
 797             { wxT("sbquo"),8218 },
 798             { wxT("scaron"),353 },
 799             { wxT("sdot"),8901 },
 800             { wxT("sect"),167 },
 801             { wxT("shy"),173 },
 802             { wxT("sigma"),963 },
 803             { wxT("sigmaf"),962 },
 804             { wxT("sim"),8764 },
 805             { wxT("spades"),9824 },
 806             { wxT("sub"),8834 },
 807             { wxT("sube"),8838 },
 808             { wxT("sum"),8721 },
 809             { wxT("sup"),8835 },
 810             { wxT("sup1"),185 },
 811             { wxT("sup2"),178 },
 812             { wxT("sup3"),179 },
 813             { wxT("supe"),8839 },
 814             { wxT("szlig"),223 },
 815             { wxT("tau"),964 },
 816             { wxT("there4"),8756 },
 817             { wxT("theta"),952 },
 818             { wxT("thetasym"),977 },
 819             { wxT("thinsp"),8201 },
 820             { wxT("thorn"),254 },
 821             { wxT("tilde"),732 },
 822             { wxT("times"),215 },
 823             { wxT("trade"),8482 },
 824             { wxT("uArr"),8657 },
 825             { wxT("uacute"),250 },
 826             { wxT("uarr"),8593 },
 827             { wxT("ucirc"),251 },
 828             { wxT("ugrave"),249 },
 829             { wxT("uml"),168 },
 830             { wxT("upsih"),978 },
 831             { wxT("upsilon"),965 },
 832             { wxT("uuml"),252 },
 833             { wxT("weierp"),8472 },
 834             { wxT("xi"),958 },
 835             { wxT("yacute"),253 },
 836             { wxT("yen"),165 },
 837             { wxT("yuml"),255 },
 838             { wxT("zeta"),950 },
 839             { wxT("zwj"),8205 },
 840             { wxT("zwnj"),8204 },
 841             {NULL, 0}};
 842         static size_t substitutions_cnt = 0;
 843
 844         if (substitutions_cnt == 0)
 845             while (substitutions[substitutions_cnt].code != 0)
 846                 substitutions_cnt++;
 847
 848         wx28HtmlEntityInfo *info = NULL;
 849 #ifdef __WXWINCE__
 850         // bsearch crashes under WinCE for some reason
 851         size_t i;
 852         for (i = 0; i < substitutions_cnt; i++)
 853         {
 854             if (entity == substitutions[i].name)
 855             {
 856                 info = & substitutions[i];
 857                 break;
 858             }
 859         }
 860 #else
 861         info = (wx28HtmlEntityInfo*) bsearch(entity.c_str(), substitutions,
 862                                            substitutions_cnt,
 863                                            sizeof(wx28HtmlEntityInfo),
 864                                            wx28HtmlEntityCompare);
 865 #endif
 866         if (info)
 867             code = info->code;
 868     }
 869
 870     if (code == 0)
 871         return 0;
 872     else
 873         return GetCharForCode(code);
 874 }
 875
 876 wxFSFile *wx28HtmlParser::OpenURL(wx28HtmlURLType WXUNUSED(type),
 877                                 const wxString& url) const
 878 {
 879     return m_FS ? m_FS->OpenFile(url) : NULL;
 880
 881 }
 882
 883
 884 //-----------------------------------------------------------------------------
 885 // wx28HtmlParser::ExtractCharsetInformation
 886 //-----------------------------------------------------------------------------
 887
 888 class wxMetaTagParser : public wx28HtmlParser
 889 {
 890 public:
 891     wxMetaTagParser() { }
 892
 893     wxObject* GetProduct() { return NULL; }
 894
 895 protected:
 896     virtual void AddText(const wxChar* WXUNUSED(txt)) {}
 897
 898     DECLARE_NO_COPY_CLASS(wxMetaTagParser)
 899 };
 900
 901 class wxMetaTagHandler : public wx28HtmlTagHandler
 902 {
 903 public:
 904     wxMetaTagHandler(wxString *retval) : wx28HtmlTagHandler(), m_retval(retval) {}
 905     wxString GetSupportedTags() { return wxT("META,BODY"); }
 906     bool HandleTag(const wx28HtmlTag& tag);
 907
 908 private:
 909     wxString *m_retval;
 910
 911     DECLARE_NO_COPY_CLASS(wxMetaTagHandler)
 912 };
 913
 914 bool wxMetaTagHandler::HandleTag(const wx28HtmlTag& tag)
 915 {
 916     if (tag.GetName() == wxT("BODY"))
 917     {
 918         m_Parser->StopParsing();
 919         return false;
 920     }
 921
 922     if (tag.HasParam(wxT("HTTP-EQUIV")) &&
 923         tag.GetParam(wxT("HTTP-EQUIV")).IsSameAs(wxT("Content-Type"), false) &&
 924         tag.HasParam(wxT("CONTENT")))
 925     {
 926         wxString content = tag.GetParam(wxT("CONTENT")).Lower();
 927         if (content.Left(19) == wxT("text/html; charset="))
 928         {
 929             *m_retval = content.Mid(19);
 930             m_Parser->StopParsing();
 931         }
 932     }
 933     return false;
 934 }
 935
 936
 937 /*static*/
 938 wxString wx28HtmlParser::ExtractCharsetInformation(const wxString& markup)
 939 {
 940     wxString charset;
 941     wxMetaTagParser *parser = new wxMetaTagParser();
 942     if(parser)
 943     {
 944         parser->AddTagHandler(new wxMetaTagHandler(&charset));
 945         parser->Parse(markup);
 946         delete parser;
 947     }
 948     return charset;
 949 }