src/html/htmlpars.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        src/html/htmlpars.cpp
   3 // Purpose:     wxHtmlParser 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 #if wxUSE_HTML && wxUSE_STREAMS
  17
  18 #ifndef WX_PRECOMP
  19     #include "wx/dynarray.h"
  20     #include "wx/log.h"
  21     #include "wx/intl.h"
  22     #include "wx/app.h"
  23     #include "wx/wxcrtvararg.h"
  24 #endif
  25
  26 #include "wx/tokenzr.h"
  27 #include "wx/wfstream.h"
  28 #include "wx/url.h"
  29 #include "wx/fontmap.h"
  30 #include "wx/html/htmldefs.h"
  31 #include "wx/html/htmlpars.h"
  32 #include "wx/arrimpl.cpp"
  33
  34 #ifdef __WXWINCE__
  35     #include "wx/msw/wince/missing.h"       // for bsearch()
  36 #endif
  37
  38 // DLL options compatibility check:
  39 WX_CHECK_BUILD_OPTIONS("wxHTML")
  40
  41 const wxChar *wxTRACE_HTML_DEBUG = _T("htmldebug");
  42
  43 //-----------------------------------------------------------------------------
  44 // wxHtmlParser helpers
  45 //-----------------------------------------------------------------------------
  46
  47 class wxHtmlTextPiece
  48 {
  49 public:
  50     wxHtmlTextPiece(int pos, int lng) : m_pos(pos), m_lng(lng) {}
  51     int m_pos, m_lng;
  52 };
  53
  54 WX_DECLARE_OBJARRAY(wxHtmlTextPiece, wxHtmlTextPieces);
  55 WX_DEFINE_OBJARRAY(wxHtmlTextPieces)
  56
  57 class wxHtmlParserState
  58 {
  59 public:
  60     wxHtmlTag         *m_curTag;
  61     wxHtmlTag         *m_tags;
  62     wxHtmlTextPieces  *m_textPieces;
  63     int                m_curTextPiece;
  64     wxString           m_source;
  65     wxHtmlParserState *m_nextState;
  66 };
  67
  68 //-----------------------------------------------------------------------------
  69 // wxHtmlParser
  70 //-----------------------------------------------------------------------------
  71
  72 IMPLEMENT_ABSTRACT_CLASS(wxHtmlParser,wxObject)
  73
  74 wxHtmlParser::wxHtmlParser()
  75     : wxObject(), m_HandlersHash(wxKEY_STRING),
  76       m_FS(NULL), m_HandlersStack(NULL)
  77 {
  78     m_entitiesParser = new wxHtmlEntitiesParser;
  79     m_Tags = NULL;
  80     m_CurTag = NULL;
  81     m_TextPieces = NULL;
  82     m_CurTextPiece = 0;
  83     m_SavedStates = NULL;
  84 }
  85
  86 wxHtmlParser::~wxHtmlParser()
  87 {
  88     while (RestoreState()) {}
  89     DestroyDOMTree();
  90
  91     if (m_HandlersStack)
  92     {
  93         wxList& tmp = *m_HandlersStack;
  94         wxList::iterator it, en;
  95         for( it = tmp.begin(), en = tmp.end(); it != en; ++it )
  96             delete (wxHashTable*)*it;
  97         tmp.clear();
  98     }
  99     delete m_HandlersStack;
 100     m_HandlersHash.Clear();
 101     WX_CLEAR_LIST(wxList, m_HandlersList);
 102     delete m_entitiesParser;
 103 }
 104
 105 wxObject* wxHtmlParser::Parse(const wxString& source)
 106 {
 107     InitParser(source);
 108     DoParsing();
 109     wxObject *result = GetProduct();
 110     DoneParser();
 111     return result;
 112 }
 113
 114 void wxHtmlParser::InitParser(const wxString& source)
 115 {
 116     SetSource(source);
 117     m_stopParsing = false;
 118 }
 119
 120 void wxHtmlParser::DoneParser()
 121 {
 122     DestroyDOMTree();
 123 }
 124
 125 void wxHtmlParser::SetSource(const wxString& src)
 126 {
 127     DestroyDOMTree();
 128     m_Source = src;
 129     CreateDOMTree();
 130     m_CurTag = NULL;
 131     m_CurTextPiece = 0;
 132 }
 133
 134 void wxHtmlParser::CreateDOMTree()
 135 {
 136     wxHtmlTagsCache cache(m_Source);
 137     m_TextPieces = new wxHtmlTextPieces;
 138     CreateDOMSubTree(NULL, 0, m_Source.length(), &cache);
 139     m_CurTextPiece = 0;
 140 }
 141
 142 extern bool wxIsCDATAElement(const wxChar *tag);
 143
 144 void wxHtmlParser::CreateDOMSubTree(wxHtmlTag *cur,
 145                                     int begin_pos, int end_pos,
 146                                     wxHtmlTagsCache *cache)
 147 {
 148     if (end_pos <= begin_pos) return;
 149
 150     wxChar c;
 151     int i = begin_pos;
 152     int textBeginning = begin_pos;
 153
 154     // If the tag contains CDATA text, we include the text between beginning
 155     // and ending tag verbosely. Setting i=end_pos will skip to the very
 156     // end of this function where text piece is added, bypassing any child
 157     // tags parsing (CDATA element can't have child elements by definition):
 158     if (cur != NULL && wxIsCDATAElement(cur->GetName().c_str()))
 159     {
 160         i = end_pos;
 161     }
 162
 163     while (i < end_pos)
 164     {
 165         c = m_Source.GetChar(i);
 166
 167         if (c == wxT('<'))
 168         {
 169             // add text to m_TextPieces:
 170             if (i - textBeginning > 0)
 171                 m_TextPieces->Add(
 172                     wxHtmlTextPiece(textBeginning, i - textBeginning));
 173
 174             // if it is a comment, skip it:
 175             wxString::const_iterator iter = m_Source.begin() + i;
 176             if ( SkipCommentTag(iter, m_Source.end()) )
 177             {
 178                 textBeginning =
 179                 i = iter - m_Source.begin() + 1; // skip closing '>' too
 180             }
 181
 182             // add another tag to the tree:
 183             else if (i < end_pos-1 && m_Source.GetChar(i+1) != wxT('/'))
 184             {
 185                 wxHtmlTag *chd;
 186                 if (cur)
 187                     chd = new wxHtmlTag(cur, m_Source,
 188                                         i, end_pos, cache, m_entitiesParser);
 189                 else
 190                 {
 191                     chd = new wxHtmlTag(NULL, m_Source,
 192                                         i, end_pos, cache, m_entitiesParser);
 193                     if (!m_Tags)
 194                     {
 195                         // if this is the first tag to be created make the root
 196                         // m_Tags point to it:
 197                         m_Tags = chd;
 198                     }
 199                     else
 200                     {
 201                         // if there is already a root tag add this tag as
 202                         // the last sibling:
 203                         chd->m_Prev = m_Tags->GetLastSibling();
 204                         chd->m_Prev->m_Next = chd;
 205                     }
 206                 }
 207
 208                 if (chd->HasEnding())
 209                 {
 210                     CreateDOMSubTree(chd,
 211                                      chd->GetBeginPos(), chd->GetEndPos1(),
 212                                      cache);
 213                     i = chd->GetEndPos2();
 214                 }
 215                 else
 216                     i = chd->GetBeginPos();
 217
 218                 textBeginning = i;
 219             }
 220
 221             // ... or skip ending tag:
 222             else
 223             {
 224                 while (i < end_pos && m_Source.GetChar(i) != wxT('>')) i++;
 225                 textBeginning = i+1;
 226             }
 227         }
 228         else i++;
 229     }
 230
 231     // add remaining text to m_TextPieces:
 232     if (end_pos - textBeginning > 0)
 233         m_TextPieces->Add(
 234             wxHtmlTextPiece(textBeginning, end_pos - textBeginning));
 235 }
 236
 237 void wxHtmlParser::DestroyDOMTree()
 238 {
 239     wxHtmlTag *t1, *t2;
 240     t1 = m_Tags;
 241     while (t1)
 242     {
 243         t2 = t1->GetNextSibling();
 244         delete t1;
 245         t1 = t2;
 246     }
 247     m_Tags = m_CurTag = NULL;
 248
 249     delete m_TextPieces;
 250     m_TextPieces = NULL;
 251 }
 252
 253 void wxHtmlParser::DoParsing()
 254 {
 255     m_CurTag = m_Tags;
 256     m_CurTextPiece = 0;
 257     DoParsing(0, m_Source.length());
 258 }
 259
 260 void wxHtmlParser::DoParsing(int begin_pos, int end_pos)
 261 {
 262     if (end_pos <= begin_pos) return;
 263
 264     wxHtmlTextPieces& pieces = *m_TextPieces;
 265     size_t piecesCnt = pieces.GetCount();
 266
 267     while (begin_pos < end_pos)
 268     {
 269         while (m_CurTag && m_CurTag->GetBeginPos() < begin_pos)
 270             m_CurTag = m_CurTag->GetNextTag();
 271         while (m_CurTextPiece < piecesCnt &&
 272                pieces[m_CurTextPiece].m_pos < begin_pos)
 273             m_CurTextPiece++;
 274
 275         if (m_CurTextPiece < piecesCnt &&
 276             (!m_CurTag ||
 277              pieces[m_CurTextPiece].m_pos < m_CurTag->GetBeginPos()))
 278         {
 279             // Add text:
 280             AddText(GetEntitiesParser()->Parse(
 281                        m_Source.Mid(pieces[m_CurTextPiece].m_pos,
 282                                     pieces[m_CurTextPiece].m_lng)));
 283             begin_pos = pieces[m_CurTextPiece].m_pos +
 284                         pieces[m_CurTextPiece].m_lng;
 285             m_CurTextPiece++;
 286         }
 287         else if (m_CurTag)
 288         {
 289             if (m_CurTag->HasEnding())
 290                 begin_pos = m_CurTag->GetEndPos2();
 291             else
 292                 begin_pos = m_CurTag->GetBeginPos();
 293             wxHtmlTag *t = m_CurTag;
 294             m_CurTag = m_CurTag->GetNextTag();
 295             AddTag(*t);
 296             if (m_stopParsing)
 297                 return;
 298         }
 299         else break;
 300     }
 301 }
 302
 303 void wxHtmlParser::AddTag(const wxHtmlTag& tag)
 304 {
 305     wxHtmlTagHandler *h;
 306     bool inner = false;
 307
 308     h = (wxHtmlTagHandler*) m_HandlersHash.Get(tag.GetName());
 309     if (h)
 310     {
 311         inner = h->HandleTag(tag);
 312         if (m_stopParsing)
 313             return;
 314     }
 315     if (!inner)
 316     {
 317         if (tag.HasEnding())
 318             DoParsing(tag.GetBeginPos(), tag.GetEndPos1());
 319     }
 320 }
 321
 322 void wxHtmlParser::AddTagHandler(wxHtmlTagHandler *handler)
 323 {
 324     wxString s(handler->GetSupportedTags());
 325     wxStringTokenizer tokenizer(s, wxT(", "));
 326
 327     while (tokenizer.HasMoreTokens())
 328         m_HandlersHash.Put(tokenizer.GetNextToken(), handler);
 329
 330     if (m_HandlersList.IndexOf(handler) == wxNOT_FOUND)
 331         m_HandlersList.Append(handler);
 332
 333     handler->SetParser(this);
 334 }
 335
 336 void wxHtmlParser::PushTagHandler(wxHtmlTagHandler *handler, const wxString& tags)
 337 {
 338     wxStringTokenizer tokenizer(tags, wxT(", "));
 339     wxString key;
 340
 341     if (m_HandlersStack == NULL)
 342     {
 343         m_HandlersStack = new wxList;
 344     }
 345
 346     m_HandlersStack->Insert((wxObject*)new wxHashTable(m_HandlersHash));
 347
 348     while (tokenizer.HasMoreTokens())
 349     {
 350         key = tokenizer.GetNextToken();
 351         m_HandlersHash.Delete(key);
 352         m_HandlersHash.Put(key, handler);
 353     }
 354 }
 355
 356 void wxHtmlParser::PopTagHandler()
 357 {
 358     wxList::compatibility_iterator first;
 359
 360     if ( !m_HandlersStack ||
 361 #if wxUSE_STL
 362          !(first = m_HandlersStack->GetFirst())
 363 #else // !wxUSE_STL
 364          ((first = m_HandlersStack->GetFirst()) == NULL)
 365 #endif // wxUSE_STL/!wxUSE_STL
 366         )
 367     {
 368         wxLogWarning(_("Warning: attempt to remove HTML tag handler from empty stack."));
 369         return;
 370     }
 371     m_HandlersHash = *((wxHashTable*) first->GetData());
 372     delete (wxHashTable*) first->GetData();
 373     m_HandlersStack->Erase(first);
 374 }
 375
 376 void wxHtmlParser::SetSourceAndSaveState(const wxString& src)
 377 {
 378     wxHtmlParserState *s = new wxHtmlParserState;
 379
 380     s->m_curTag = m_CurTag;
 381     s->m_tags = m_Tags;
 382     s->m_textPieces = m_TextPieces;
 383     s->m_curTextPiece = m_CurTextPiece;
 384     s->m_source = m_Source;
 385
 386     s->m_nextState = m_SavedStates;
 387     m_SavedStates = s;
 388
 389     m_CurTag = NULL;
 390     m_Tags = NULL;
 391     m_TextPieces = NULL;
 392     m_CurTextPiece = 0;
 393     m_Source = wxEmptyString;
 394
 395     SetSource(src);
 396 }
 397
 398 bool wxHtmlParser::RestoreState()
 399 {
 400     if (!m_SavedStates) return false;
 401
 402     DestroyDOMTree();
 403
 404     wxHtmlParserState *s = m_SavedStates;
 405     m_SavedStates = s->m_nextState;
 406
 407     m_CurTag = s->m_curTag;
 408     m_Tags = s->m_tags;
 409     m_TextPieces = s->m_textPieces;
 410     m_CurTextPiece = s->m_curTextPiece;
 411     m_Source = s->m_source;
 412
 413     delete s;
 414     return true;
 415 }
 416
 417 wxString wxHtmlParser::GetInnerSource(const wxHtmlTag& tag)
 418 {
 419     return GetSource()->Mid(tag.GetBeginPos(),
 420                             tag.GetEndPos1() - tag.GetBeginPos());
 421 }
 422
 423 //-----------------------------------------------------------------------------
 424 // wxHtmlTagHandler
 425 //-----------------------------------------------------------------------------
 426
 427 IMPLEMENT_ABSTRACT_CLASS(wxHtmlTagHandler,wxObject)
 428
 429 void wxHtmlTagHandler::ParseInnerSource(const wxString& source)
 430 {
 431     // It is safe to temporarily change the source being parsed,
 432     // provided we restore the state back after parsing
 433     m_Parser->SetSourceAndSaveState(source);
 434     m_Parser->DoParsing();
 435     m_Parser->RestoreState();
 436 }
 437
 438
 439 //-----------------------------------------------------------------------------
 440 // wxHtmlEntitiesParser
 441 //-----------------------------------------------------------------------------
 442
 443 IMPLEMENT_DYNAMIC_CLASS(wxHtmlEntitiesParser,wxObject)
 444
 445 wxHtmlEntitiesParser::wxHtmlEntitiesParser()
 446 #if wxUSE_WCHAR_T && !wxUSE_UNICODE
 447     : m_conv(NULL), m_encoding(wxFONTENCODING_SYSTEM)
 448 #endif
 449 {
 450 }
 451
 452 wxHtmlEntitiesParser::~wxHtmlEntitiesParser()
 453 {
 454 #if wxUSE_WCHAR_T && !wxUSE_UNICODE
 455     delete m_conv;
 456 #endif
 457 }
 458
 459 void wxHtmlEntitiesParser::SetEncoding(wxFontEncoding encoding)
 460 {
 461 #if wxUSE_WCHAR_T && !wxUSE_UNICODE
 462     if (encoding == m_encoding)
 463         return;
 464
 465     delete m_conv;
 466
 467     m_encoding = encoding;
 468     if (m_encoding == wxFONTENCODING_SYSTEM)
 469         m_conv = NULL;
 470     else
 471         m_conv = new wxCSConv(wxFontMapper::GetEncodingName(m_encoding));
 472 #else
 473     (void) encoding;
 474 #endif
 475 }
 476
 477 wxString wxHtmlEntitiesParser::Parse(const wxString& input)
 478 {
 479     const wxChar *c, *last;
 480     const wxChar *in_str = input.c_str();
 481     wxString output;
 482
 483     output.reserve(input.length());
 484
 485     for (c = in_str, last = in_str; *c != wxT('\0'); c++)
 486     {
 487         if (*c == wxT('&'))
 488         {
 489             if (c - last > 0)
 490                 output.append(last, c - last);
 491             if ( *++c == wxT('\0') )
 492                 break;
 493
 494             wxString entity;
 495             const wxChar *ent_s = c;
 496             wxChar entity_char;
 497
 498             for (; (*c >= wxT('a') && *c <= wxT('z')) ||
 499                    (*c >= wxT('A') && *c <= wxT('Z')) ||
 500                    (*c >= wxT('0') && *c <= wxT('9')) ||
 501                    *c == wxT('_') || *c == wxT('#'); c++) {}
 502             entity.append(ent_s, c - ent_s);
 503             if (*c != wxT(';')) c--;
 504             last = c+1;
 505             entity_char = GetEntityChar(entity);
 506             if (entity_char)
 507                 output << entity_char;
 508             else
 509             {
 510                 output.append(ent_s-1, c-ent_s+2);
 511                 wxLogTrace(wxTRACE_HTML_DEBUG,
 512                            wxT("Unrecognized HTML entity: '%s'"),
 513                            entity.c_str());
 514             }
 515         }
 516     }
 517     if (*last != wxT('\0'))
 518         output.append(last);
 519     return output;
 520 }
 521
 522 struct wxHtmlEntityInfo
 523 {
 524     const wxChar *name;
 525     unsigned code;
 526 };
 527
 528 extern "C" int LINKAGEMODE wxHtmlEntityCompare(const void *key, const void *item)
 529 {
 530     return wxStrcmp((wxChar*)key, ((wxHtmlEntityInfo*)item)->name);
 531 }
 532
 533 #if !wxUSE_UNICODE
 534 wxChar wxHtmlEntitiesParser::GetCharForCode(unsigned code)
 535 {
 536 #if wxUSE_WCHAR_T
 537     char buf[2];
 538     wchar_t wbuf[2];
 539     wbuf[0] = (wchar_t)code;
 540     wbuf[1] = 0;
 541     wxMBConv *conv = m_conv ? m_conv : &wxConvLocal;
 542     if (conv->WC2MB(buf, wbuf, 2) == (size_t)-1)
 543         return '?';
 544     return buf[0];
 545 #else
 546     return (code < 256) ? (wxChar)code : '?';
 547 #endif
 548 }
 549 #endif
 550
 551 wxChar wxHtmlEntitiesParser::GetEntityChar(const wxString& entity)
 552 {
 553     unsigned code = 0;
 554
 555     if (entity[0] == wxT('#'))
 556     {
 557         const wxChar *ent_s = entity.c_str();
 558         const wxChar *format;
 559
 560         if (ent_s[1] == wxT('x') || ent_s[1] == wxT('X'))
 561         {
 562             format = wxT("%x");
 563             ent_s++;
 564         }
 565         else
 566             format = wxT("%u");
 567         ent_s++;
 568
 569         if (wxSscanf(ent_s, format, &code) != 1)
 570             code = 0;
 571     }
 572     else
 573     {
 574         static wxHtmlEntityInfo substitutions[] = {
 575             { wxT("AElig"),198 },
 576             { wxT("Aacute"),193 },
 577             { wxT("Acirc"),194 },
 578             { wxT("Agrave"),192 },
 579             { wxT("Alpha"),913 },
 580             { wxT("Aring"),197 },
 581             { wxT("Atilde"),195 },
 582             { wxT("Auml"),196 },
 583             { wxT("Beta"),914 },
 584             { wxT("Ccedil"),199 },
 585             { wxT("Chi"),935 },
 586             { wxT("Dagger"),8225 },
 587             { wxT("Delta"),916 },
 588             { wxT("ETH"),208 },
 589             { wxT("Eacute"),201 },
 590             { wxT("Ecirc"),202 },
 591             { wxT("Egrave"),200 },
 592             { wxT("Epsilon"),917 },
 593             { wxT("Eta"),919 },
 594             { wxT("Euml"),203 },
 595             { wxT("Gamma"),915 },
 596             { wxT("Iacute"),205 },
 597             { wxT("Icirc"),206 },
 598             { wxT("Igrave"),204 },
 599             { wxT("Iota"),921 },
 600             { wxT("Iuml"),207 },
 601             { wxT("Kappa"),922 },
 602             { wxT("Lambda"),923 },
 603             { wxT("Mu"),924 },
 604             { wxT("Ntilde"),209 },
 605             { wxT("Nu"),925 },
 606             { wxT("OElig"),338 },
 607             { wxT("Oacute"),211 },
 608             { wxT("Ocirc"),212 },
 609             { wxT("Ograve"),210 },
 610             { wxT("Omega"),937 },
 611             { wxT("Omicron"),927 },
 612             { wxT("Oslash"),216 },
 613             { wxT("Otilde"),213 },
 614             { wxT("Ouml"),214 },
 615             { wxT("Phi"),934 },
 616             { wxT("Pi"),928 },
 617             { wxT("Prime"),8243 },
 618             { wxT("Psi"),936 },
 619             { wxT("Rho"),929 },
 620             { wxT("Scaron"),352 },
 621             { wxT("Sigma"),931 },
 622             { wxT("THORN"),222 },
 623             { wxT("Tau"),932 },
 624             { wxT("Theta"),920 },
 625             { wxT("Uacute"),218 },
 626             { wxT("Ucirc"),219 },
 627             { wxT("Ugrave"),217 },
 628             { wxT("Upsilon"),933 },
 629             { wxT("Uuml"),220 },
 630             { wxT("Xi"),926 },
 631             { wxT("Yacute"),221 },
 632             { wxT("Yuml"),376 },
 633             { wxT("Zeta"),918 },
 634             { wxT("aacute"),225 },
 635             { wxT("acirc"),226 },
 636             { wxT("acute"),180 },
 637             { wxT("aelig"),230 },
 638             { wxT("agrave"),224 },
 639             { wxT("alefsym"),8501 },
 640             { wxT("alpha"),945 },
 641             { wxT("amp"),38 },
 642             { wxT("and"),8743 },
 643             { wxT("ang"),8736 },
 644             { wxT("aring"),229 },
 645             { wxT("asymp"),8776 },
 646             { wxT("atilde"),227 },
 647             { wxT("auml"),228 },
 648             { wxT("bdquo"),8222 },
 649             { wxT("beta"),946 },
 650             { wxT("brvbar"),166 },
 651             { wxT("bull"),8226 },
 652             { wxT("cap"),8745 },
 653             { wxT("ccedil"),231 },
 654             { wxT("cedil"),184 },
 655             { wxT("cent"),162 },
 656             { wxT("chi"),967 },
 657             { wxT("circ"),710 },
 658             { wxT("clubs"),9827 },
 659             { wxT("cong"),8773 },
 660             { wxT("copy"),169 },
 661             { wxT("crarr"),8629 },
 662             { wxT("cup"),8746 },
 663             { wxT("curren"),164 },
 664             { wxT("dArr"),8659 },
 665             { wxT("dagger"),8224 },
 666             { wxT("darr"),8595 },
 667             { wxT("deg"),176 },
 668             { wxT("delta"),948 },
 669             { wxT("diams"),9830 },
 670             { wxT("divide"),247 },
 671             { wxT("eacute"),233 },
 672             { wxT("ecirc"),234 },
 673             { wxT("egrave"),232 },
 674             { wxT("empty"),8709 },
 675             { wxT("emsp"),8195 },
 676             { wxT("ensp"),8194 },
 677             { wxT("epsilon"),949 },
 678             { wxT("equiv"),8801 },
 679             { wxT("eta"),951 },
 680             { wxT("eth"),240 },
 681             { wxT("euml"),235 },
 682             { wxT("euro"),8364 },
 683             { wxT("exist"),8707 },
 684             { wxT("fnof"),402 },
 685             { wxT("forall"),8704 },
 686             { wxT("frac12"),189 },
 687             { wxT("frac14"),188 },
 688             { wxT("frac34"),190 },
 689             { wxT("frasl"),8260 },
 690             { wxT("gamma"),947 },
 691             { wxT("ge"),8805 },
 692             { wxT("gt"),62 },
 693             { wxT("hArr"),8660 },
 694             { wxT("harr"),8596 },
 695             { wxT("hearts"),9829 },
 696             { wxT("hellip"),8230 },
 697             { wxT("iacute"),237 },
 698             { wxT("icirc"),238 },
 699             { wxT("iexcl"),161 },
 700             { wxT("igrave"),236 },
 701             { wxT("image"),8465 },
 702             { wxT("infin"),8734 },
 703             { wxT("int"),8747 },
 704             { wxT("iota"),953 },
 705             { wxT("iquest"),191 },
 706             { wxT("isin"),8712 },
 707             { wxT("iuml"),239 },
 708             { wxT("kappa"),954 },
 709             { wxT("lArr"),8656 },
 710             { wxT("lambda"),955 },
 711             { wxT("lang"),9001 },
 712             { wxT("laquo"),171 },
 713             { wxT("larr"),8592 },
 714             { wxT("lceil"),8968 },
 715             { wxT("ldquo"),8220 },
 716             { wxT("le"),8804 },
 717             { wxT("lfloor"),8970 },
 718             { wxT("lowast"),8727 },
 719             { wxT("loz"),9674 },
 720             { wxT("lrm"),8206 },
 721             { wxT("lsaquo"),8249 },
 722             { wxT("lsquo"),8216 },
 723             { wxT("lt"),60 },
 724             { wxT("macr"),175 },
 725             { wxT("mdash"),8212 },
 726             { wxT("micro"),181 },
 727             { wxT("middot"),183 },
 728             { wxT("minus"),8722 },
 729             { wxT("mu"),956 },
 730             { wxT("nabla"),8711 },
 731             { wxT("nbsp"),160 },
 732             { wxT("ndash"),8211 },
 733             { wxT("ne"),8800 },
 734             { wxT("ni"),8715 },
 735             { wxT("not"),172 },
 736             { wxT("notin"),8713 },
 737             { wxT("nsub"),8836 },
 738             { wxT("ntilde"),241 },
 739             { wxT("nu"),957 },
 740             { wxT("oacute"),243 },
 741             { wxT("ocirc"),244 },
 742             { wxT("oelig"),339 },
 743             { wxT("ograve"),242 },
 744             { wxT("oline"),8254 },
 745             { wxT("omega"),969 },
 746             { wxT("omicron"),959 },
 747             { wxT("oplus"),8853 },
 748             { wxT("or"),8744 },
 749             { wxT("ordf"),170 },
 750             { wxT("ordm"),186 },
 751             { wxT("oslash"),248 },
 752             { wxT("otilde"),245 },
 753             { wxT("otimes"),8855 },
 754             { wxT("ouml"),246 },
 755             { wxT("para"),182 },
 756             { wxT("part"),8706 },
 757             { wxT("permil"),8240 },
 758             { wxT("perp"),8869 },
 759             { wxT("phi"),966 },
 760             { wxT("pi"),960 },
 761             { wxT("piv"),982 },
 762             { wxT("plusmn"),177 },
 763             { wxT("pound"),163 },
 764             { wxT("prime"),8242 },
 765             { wxT("prod"),8719 },
 766             { wxT("prop"),8733 },
 767             { wxT("psi"),968 },
 768             { wxT("quot"),34 },
 769             { wxT("rArr"),8658 },
 770             { wxT("radic"),8730 },
 771             { wxT("rang"),9002 },
 772             { wxT("raquo"),187 },
 773             { wxT("rarr"),8594 },
 774             { wxT("rceil"),8969 },
 775             { wxT("rdquo"),8221 },
 776             { wxT("real"),8476 },
 777             { wxT("reg"),174 },
 778             { wxT("rfloor"),8971 },
 779             { wxT("rho"),961 },
 780             { wxT("rlm"),8207 },
 781             { wxT("rsaquo"),8250 },
 782             { wxT("rsquo"),8217 },
 783             { wxT("sbquo"),8218 },
 784             { wxT("scaron"),353 },
 785             { wxT("sdot"),8901 },
 786             { wxT("sect"),167 },
 787             { wxT("shy"),173 },
 788             { wxT("sigma"),963 },
 789             { wxT("sigmaf"),962 },
 790             { wxT("sim"),8764 },
 791             { wxT("spades"),9824 },
 792             { wxT("sub"),8834 },
 793             { wxT("sube"),8838 },
 794             { wxT("sum"),8721 },
 795             { wxT("sup"),8835 },
 796             { wxT("sup1"),185 },
 797             { wxT("sup2"),178 },
 798             { wxT("sup3"),179 },
 799             { wxT("supe"),8839 },
 800             { wxT("szlig"),223 },
 801             { wxT("tau"),964 },
 802             { wxT("there4"),8756 },
 803             { wxT("theta"),952 },
 804             { wxT("thetasym"),977 },
 805             { wxT("thinsp"),8201 },
 806             { wxT("thorn"),254 },
 807             { wxT("tilde"),732 },
 808             { wxT("times"),215 },
 809             { wxT("trade"),8482 },
 810             { wxT("uArr"),8657 },
 811             { wxT("uacute"),250 },
 812             { wxT("uarr"),8593 },
 813             { wxT("ucirc"),251 },
 814             { wxT("ugrave"),249 },
 815             { wxT("uml"),168 },
 816             { wxT("upsih"),978 },
 817             { wxT("upsilon"),965 },
 818             { wxT("uuml"),252 },
 819             { wxT("weierp"),8472 },
 820             { wxT("xi"),958 },
 821             { wxT("yacute"),253 },
 822             { wxT("yen"),165 },
 823             { wxT("yuml"),255 },
 824             { wxT("zeta"),950 },
 825             { wxT("zwj"),8205 },
 826             { wxT("zwnj"),8204 },
 827             {NULL, 0}};
 828         static size_t substitutions_cnt = 0;
 829
 830         if (substitutions_cnt == 0)
 831             while (substitutions[substitutions_cnt].code != 0)
 832                 substitutions_cnt++;
 833
 834         wxHtmlEntityInfo *info = NULL;
 835 #ifdef __WXWINCE__
 836         // bsearch crashes under WinCE for some reason
 837         size_t i;
 838         for (i = 0; i < substitutions_cnt; i++)
 839         {
 840             if (entity == substitutions[i].name)
 841             {
 842                 info = & substitutions[i];
 843                 break;
 844             }
 845         }
 846 #else
 847         info = (wxHtmlEntityInfo*) bsearch(entity.c_str(), substitutions,
 848                                            substitutions_cnt,
 849                                            sizeof(wxHtmlEntityInfo),
 850                                            wxHtmlEntityCompare);
 851 #endif
 852         if (info)
 853             code = info->code;
 854     }
 855
 856     if (code == 0)
 857         return 0;
 858     else
 859         return GetCharForCode(code);
 860 }
 861
 862 wxFSFile *wxHtmlParser::OpenURL(wxHtmlURLType WXUNUSED(type),
 863                                 const wxString& url) const
 864 {
 865     return m_FS ? m_FS->OpenFile(url) : NULL;
 866
 867 }
 868
 869
 870 //-----------------------------------------------------------------------------
 871 // wxHtmlParser::ExtractCharsetInformation
 872 //-----------------------------------------------------------------------------
 873
 874 class wxMetaTagParser : public wxHtmlParser
 875 {
 876 public:
 877     wxMetaTagParser() { }
 878
 879     wxObject* GetProduct() { return NULL; }
 880
 881 protected:
 882     virtual void AddText(const wxChar* WXUNUSED(txt)) {}
 883
 884     DECLARE_NO_COPY_CLASS(wxMetaTagParser)
 885 };
 886
 887 class wxMetaTagHandler : public wxHtmlTagHandler
 888 {
 889 public:
 890     wxMetaTagHandler(wxString *retval) : wxHtmlTagHandler(), m_retval(retval) {}
 891     wxString GetSupportedTags() { return wxT("META,BODY"); }
 892     bool HandleTag(const wxHtmlTag& tag);
 893
 894 private:
 895     wxString *m_retval;
 896
 897     DECLARE_NO_COPY_CLASS(wxMetaTagHandler)
 898 };
 899
 900 bool wxMetaTagHandler::HandleTag(const wxHtmlTag& tag)
 901 {
 902     if (tag.GetName() == _T("BODY"))
 903     {
 904         m_Parser->StopParsing();
 905         return false;
 906     }
 907
 908     if (tag.HasParam(_T("HTTP-EQUIV")) &&
 909         tag.GetParam(_T("HTTP-EQUIV")).IsSameAs(_T("Content-Type"), false) &&
 910         tag.HasParam(_T("CONTENT")))
 911     {
 912         wxString content = tag.GetParam(_T("CONTENT")).Lower();
 913         if (content.Left(19) == _T("text/html; charset="))
 914         {
 915             *m_retval = content.Mid(19);
 916             m_Parser->StopParsing();
 917         }
 918     }
 919     return false;
 920 }
 921
 922
 923 /*static*/
 924 wxString wxHtmlParser::ExtractCharsetInformation(const wxString& markup)
 925 {
 926     wxString charset;
 927     wxMetaTagParser *parser = new wxMetaTagParser();
 928     if(parser)
 929     {
 930         parser->AddTagHandler(new wxMetaTagHandler(&charset));
 931         parser->Parse(markup);
 932         delete parser;
 933     }
 934     return charset;
 935 }
 936
 937 /* static */
 938 bool
 939 wxHtmlParser::SkipCommentTag(wxString::const_iterator& start,
 940                              wxString::const_iterator end)
 941 {
 942     wxASSERT_MSG( *start == '<', _T("should be called on the tag start") );
 943
 944     wxString::const_iterator p = start;
 945
 946     // comments begin with "<!--" in HTML 4.0
 947     if ( end - p < 3 || *++p != '!' || *++p != '-' || *++p != '-' )
 948     {
 949         // not a comment at all
 950         return false;
 951     }
 952
 953     // skip the start of the comment tag in any case, if we don't find the
 954     // closing tag we should ignore broken markup
 955     start = p;
 956
 957     // comments end with "--[ \t\r\n]*>", i.e. white space is allowed between
 958     // comment delimiter and the closing tag character (section 3.2.4 of
 959     // http://www.w3.org/TR/html401/)
 960     int dashes = 0;
 961     while ( ++p < end )
 962     {
 963         const wxChar c = *p;
 964
 965         if ( (c == wxT(' ') || c == wxT('\n') ||
 966               c == wxT('\r') || c == wxT('\t')) && dashes >= 2 )
 967         {
 968             // ignore white space before potential tag end
 969             continue;
 970         }
 971
 972         if ( c == wxT('>') && dashes >= 2 )
 973         {
 974             // found end of comment
 975             start = p;
 976             break;
 977         }
 978
 979         if ( c == wxT('-') )
 980             dashes++;
 981         else
 982             dashes = 0;
 983     }
 984
 985     return true;
 986 }
 987
 988 #endif // wxUSE_HTML