tests/benchmarks/htmlparser/htmlpars.cpp

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