src/html/htmltag.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        src/html/htmltag.cpp
   3 // Purpose:     wxHtmlTag class (represents single tag)
   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
  17
  18 #include "wx/html/htmltag.h"
  19
  20 #ifndef WX_PRECOMP
  21     #include "wx/colour.h"
  22     #include "wx/wxcrtvararg.h"
  23 #endif
  24
  25 #include "wx/html/htmlpars.h"
  26 #include "wx/html/styleparams.h"
  27
  28 #include "wx/vector.h"
  29
  30 #include <stdio.h> // for vsscanf
  31 #include <stdarg.h>
  32
  33 //-----------------------------------------------------------------------------
  34 // wxHtmlTagsCache
  35 //-----------------------------------------------------------------------------
  36
  37 struct wxHtmlCacheItem
  38 {
  39     // this is "pos" value passed to wxHtmlTag's constructor.
  40     // it is position of '<' character of the tag
  41     wxString::const_iterator Key;
  42
  43     // Tag type
  44     enum Type
  45     {
  46         Type_Normal, // normal tag with a matching ending tag
  47         Type_NoMatchingEndingTag, // there's no ending tag for this tag
  48         Type_EndingTag // this is ending tag </..>
  49     };
  50     Type type;
  51
  52     // end positions for the tag:
  53     // end1 is '<' of ending tag,
  54     // end2 is '>' or both are
  55     wxString::const_iterator End1, End2;
  56
  57     // name of this tag
  58     wxChar *Name;
  59 };
  60
  61 // NB: this is an empty class and not typedef because of forward declaration
  62 class wxHtmlTagsCacheData : public wxVector<wxHtmlCacheItem>
  63 {
  64 };
  65
  66 bool wxIsCDATAElement(const wxChar *tag)
  67 {
  68     return (wxStrcmp(tag, wxT("SCRIPT")) == 0) ||
  69            (wxStrcmp(tag, wxT("STYLE")) == 0);
  70 }
  71
  72 bool wxIsCDATAElement(const wxString& tag)
  73 {
  74     return (wxStrcmp(tag.wx_str(), wxS("SCRIPT")) == 0) ||
  75            (wxStrcmp(tag.wx_str(), wxS("STYLE")) == 0);
  76 }
  77
  78 wxHtmlTagsCache::wxHtmlTagsCache(const wxString& source)
  79 {
  80     m_Cache = new wxHtmlTagsCacheData;
  81     m_CachePos = 0;
  82
  83     wxChar tagBuffer[256];
  84
  85     const wxString::const_iterator end = source.end();
  86     for ( wxString::const_iterator pos = source.begin(); pos < end; ++pos )
  87     {
  88         if (*pos == wxT('<'))   // tag found:
  89         {
  90             // don't cache comment tags
  91             if ( wxHtmlParser::SkipCommentTag(pos, source.end()) )
  92                 continue;
  93
  94             size_t tg = Cache().size();
  95             Cache().push_back(wxHtmlCacheItem());
  96
  97             wxString::const_iterator stpos = pos++;
  98             Cache()[tg].Key = stpos;
  99
 100             int i;
 101             for ( i = 0;
 102                   pos < end && i < (int)WXSIZEOF(tagBuffer) - 1 &&
 103                   *pos != wxT('>') && !wxIsspace(*pos);
 104                   ++i, ++pos )
 105             {
 106                 tagBuffer[i] = (wxChar)wxToupper(*pos);
 107             }
 108             tagBuffer[i] = wxT('\0');
 109
 110             Cache()[tg].Name = new wxChar[i+1];
 111             memcpy(Cache()[tg].Name, tagBuffer, (i+1)*sizeof(wxChar));
 112
 113             while (pos < end && *pos != wxT('>'))
 114                 ++pos;
 115
 116             if ((stpos+1) < end && *(stpos+1) == wxT('/')) // ending tag:
 117             {
 118                 Cache()[tg].type = wxHtmlCacheItem::Type_EndingTag;
 119                 // find matching begin tag:
 120                 for (i = tg; i >= 0; i--)
 121                 {
 122                     if ((Cache()[i].type == wxHtmlCacheItem::Type_NoMatchingEndingTag) && (wxStrcmp(Cache()[i].Name, tagBuffer+1) == 0))
 123                     {
 124                         Cache()[i].type = wxHtmlCacheItem::Type_Normal;
 125                         Cache()[i].End1 = stpos;
 126                         Cache()[i].End2 = pos + 1;
 127                         break;
 128                     }
 129                 }
 130             }
 131             else
 132             {
 133                 Cache()[tg].type = wxHtmlCacheItem::Type_NoMatchingEndingTag;
 134
 135                 if (wxIsCDATAElement(tagBuffer))
 136                 {
 137                     // store the orig pos in case we are missing the closing
 138                     // tag (see below)
 139                     const wxString::const_iterator old_pos = pos;
 140                     bool foundCloseTag = false;
 141
 142                     // find next matching tag
 143                     int tag_len = wxStrlen(tagBuffer);
 144                     while (pos < end)
 145                     {
 146                         // find the ending tag
 147                         while (pos + 1 < end &&
 148                                (*pos != '<' || *(pos+1) != '/'))
 149                             ++pos;
 150                         if (*pos == '<')
 151                             ++pos;
 152
 153                         // see if it matches
 154                         int match_pos = 0;
 155                         while (pos < end && match_pos < tag_len )
 156                         {
 157                             wxChar c = *pos;
 158                             if ( c == '>' || c == '<' )
 159                                 break;
 160
 161                             // cast to wxChar needed to suppress warning in
 162                             // Unicode build
 163                             if ((wxChar)wxToupper(c) == tagBuffer[match_pos])
 164                             {
 165                                 ++match_pos;
 166                             }
 167                             else if (c == wxT(' ') || c == wxT('\n') ||
 168                                 c == wxT('\r') || c == wxT('\t'))
 169                             {
 170                                 // need to skip over these
 171                             }
 172                             else
 173                             {
 174                                 match_pos = 0;
 175                             }
 176                             ++pos;
 177                         }
 178
 179                         // found a match
 180                         if (match_pos == tag_len)
 181                         {
 182                             pos = pos - tag_len - 3;
 183                             foundCloseTag = true;
 184                             break;
 185                         }
 186                         else // keep looking for the closing tag
 187                         {
 188                             ++pos;
 189                         }
 190                     }
 191                     if (!foundCloseTag)
 192                     {
 193                         // we didn't find closing tag; this means the markup
 194                         // is incorrect and the best thing we can do is to
 195                         // ignore the unclosed tag and continue parsing as if
 196                         // it didn't exist:
 197                         pos = old_pos;
 198                     }
 199                 }
 200             }
 201         }
 202     }
 203
 204     // ok, we're done, now we'll free .Name members of cache - we don't need it anymore:
 205     for ( wxHtmlTagsCacheData::iterator i = Cache().begin();
 206           i != Cache().end(); ++i )
 207     {
 208         delete[] i->Name;
 209         i->Name = NULL;
 210     }
 211 }
 212
 213 wxHtmlTagsCache::~wxHtmlTagsCache()
 214 {
 215     delete m_Cache;
 216 }
 217
 218 void wxHtmlTagsCache::QueryTag(const wxString::const_iterator& at,
 219                                const wxString::const_iterator& inputEnd,
 220                                wxString::const_iterator *end1,
 221                                wxString::const_iterator *end2,
 222                                bool *hasEnding)
 223 {
 224     if (Cache().empty())
 225         return;
 226
 227     if (Cache()[m_CachePos].Key != at)
 228     {
 229         int delta = (at < Cache()[m_CachePos].Key) ? -1 : 1;
 230         do
 231         {
 232             m_CachePos += delta;
 233
 234             if ( m_CachePos < 0 || m_CachePos >= (int)Cache().size() )
 235             {
 236                 if ( m_CachePos < 0 )
 237                     m_CachePos = 0;
 238                 else
 239                     m_CachePos = Cache().size() - 1;
 240                 // something is very wrong with HTML, give up by returning an
 241                 // impossibly large value which is going to be ignored by the
 242                 // caller
 243                 *end1 =
 244                 *end2 = inputEnd;
 245                 *hasEnding = true;
 246                 return;
 247             }
 248         }
 249         while (Cache()[m_CachePos].Key != at);
 250     }
 251
 252     switch ( Cache()[m_CachePos].type )
 253     {
 254         case wxHtmlCacheItem::Type_Normal:
 255             *end1 = Cache()[m_CachePos].End1;
 256             *end2 = Cache()[m_CachePos].End2;
 257             *hasEnding = true;
 258             break;
 259
 260         case wxHtmlCacheItem::Type_EndingTag:
 261             wxFAIL_MSG("QueryTag called for ending tag - can't be");
 262             // but if it does happen, fall through, better than crashing
 263
 264         case wxHtmlCacheItem::Type_NoMatchingEndingTag:
 265             // If input HTML is invalid and there's no closing tag for this
 266             // one, pretend that it runs all the way to the end of input
 267             *end1 = inputEnd;
 268             *end2 = inputEnd;
 269             *hasEnding = false;
 270             break;
 271     }
 272 }
 273
 274
 275
 276
 277 //-----------------------------------------------------------------------------
 278 // wxHtmlTag
 279 //-----------------------------------------------------------------------------
 280
 281 wxHtmlTag::wxHtmlTag(wxHtmlTag *parent,
 282                      const wxString *source,
 283                      const wxString::const_iterator& pos,
 284                      const wxString::const_iterator& end_pos,
 285                      wxHtmlTagsCache *cache,
 286                      wxHtmlEntitiesParser *entParser)
 287 {
 288     /* Setup DOM relations */
 289
 290     m_Next = NULL;
 291     m_FirstChild = m_LastChild = NULL;
 292     m_Parent = parent;
 293     if (parent)
 294     {
 295         m_Prev = m_Parent->m_LastChild;
 296         if (m_Prev == NULL)
 297             m_Parent->m_FirstChild = this;
 298         else
 299             m_Prev->m_Next = this;
 300         m_Parent->m_LastChild = this;
 301     }
 302     else
 303         m_Prev = NULL;
 304
 305     /* Find parameters and their values: */
 306
 307     wxChar c wxDUMMY_INITIALIZE(0);
 308
 309     // fill-in name, params and begin pos:
 310     wxString::const_iterator i(pos+1);
 311
 312     // find tag's name and convert it to uppercase:
 313     while ((i < end_pos) &&
 314            ((c = *(i++)) != wxT(' ') && c != wxT('\r') &&
 315              c != wxT('\n') && c != wxT('\t') &&
 316              c != wxT('>') && c != wxT('/')))
 317     {
 318         if ((c >= wxT('a')) && (c <= wxT('z')))
 319             c -= (wxT('a') - wxT('A'));
 320         m_Name << c;
 321     }
 322
 323     // if the tag has parameters, read them and "normalize" them,
 324     // i.e. convert to uppercase, replace whitespaces by spaces and
 325     // remove whitespaces around '=':
 326     if (*(i-1) != wxT('>'))
 327     {
 328         #define IS_WHITE(c) (c == wxT(' ') || c == wxT('\r') || \
 329                              c == wxT('\n') || c == wxT('\t'))
 330         wxString pname, pvalue;
 331         wxChar quote;
 332         enum
 333         {
 334             ST_BEFORE_NAME = 1,
 335             ST_NAME,
 336             ST_BEFORE_EQ,
 337             ST_BEFORE_VALUE,
 338             ST_VALUE
 339         } state;
 340
 341         quote = 0;
 342         state = ST_BEFORE_NAME;
 343         while (i < end_pos)
 344         {
 345             c = *(i++);
 346
 347             if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
 348             {
 349                 if (state == ST_BEFORE_EQ || state == ST_NAME)
 350                 {
 351                     m_ParamNames.Add(pname);
 352                     m_ParamValues.Add(wxGetEmptyString());
 353                 }
 354                 else if (state == ST_VALUE && quote == 0)
 355                 {
 356                     m_ParamNames.Add(pname);
 357                     if (entParser)
 358                         m_ParamValues.Add(entParser->Parse(pvalue));
 359                     else
 360                         m_ParamValues.Add(pvalue);
 361                 }
 362                 break;
 363             }
 364             switch (state)
 365             {
 366                 case ST_BEFORE_NAME:
 367                     if (!IS_WHITE(c))
 368                     {
 369                         pname = c;
 370                         state = ST_NAME;
 371                     }
 372                     break;
 373                 case ST_NAME:
 374                     if (IS_WHITE(c))
 375                         state = ST_BEFORE_EQ;
 376                     else if (c == wxT('='))
 377                         state = ST_BEFORE_VALUE;
 378                     else
 379                         pname << c;
 380                     break;
 381                 case ST_BEFORE_EQ:
 382                     if (c == wxT('='))
 383                         state = ST_BEFORE_VALUE;
 384                     else if (!IS_WHITE(c))
 385                     {
 386                         m_ParamNames.Add(pname);
 387                         m_ParamValues.Add(wxGetEmptyString());
 388                         pname = c;
 389                         state = ST_NAME;
 390                     }
 391                     break;
 392                 case ST_BEFORE_VALUE:
 393                     if (!IS_WHITE(c))
 394                     {
 395                         if (c == wxT('"') || c == wxT('\''))
 396                             quote = c, pvalue = wxGetEmptyString();
 397                         else
 398                             quote = 0, pvalue = c;
 399                         state = ST_VALUE;
 400                     }
 401                     break;
 402                 case ST_VALUE:
 403                     if ((quote != 0 && c == quote) ||
 404                         (quote == 0 && IS_WHITE(c)))
 405                     {
 406                         m_ParamNames.Add(pname);
 407                         if (quote == 0)
 408                         {
 409                             // VS: backward compatibility, no real reason,
 410                             //     but wxHTML code relies on this... :(
 411                             pvalue.MakeUpper();
 412                         }
 413                         if (entParser)
 414                             m_ParamValues.Add(entParser->Parse(pvalue));
 415                         else
 416                             m_ParamValues.Add(pvalue);
 417                         state = ST_BEFORE_NAME;
 418                     }
 419                     else
 420                         pvalue << c;
 421                     break;
 422             }
 423         }
 424
 425         #undef IS_WHITE
 426     }
 427     m_Begin = i;
 428     cache->QueryTag(pos, source->end(), &m_End1, &m_End2, &m_hasEnding);
 429     if (m_End1 > end_pos) m_End1 = end_pos;
 430     if (m_End2 > end_pos) m_End2 = end_pos;
 431
 432 #if WXWIN_COMPATIBILITY_2_8
 433     m_sourceStart = source->begin();
 434 #endif
 435
 436     // Try to parse any style parameters that can be handled simply by
 437     // converting them to the equivalent HTML 3 attributes: this is a far cry
 438     // from perfect but better than nothing.
 439     static const struct EquivAttr
 440     {
 441         const char *style;
 442         const char *attr;
 443     } equivAttrs[] =
 444     {
 445         { "text-align",         "ALIGN"         },
 446         { "width",              "WIDTH"         },
 447         { "vertical-align",     "VALIGN"        },
 448         { "background",         "BGCOLOR"       },
 449     };
 450
 451     wxHtmlStyleParams styleParams(*this);
 452     for ( unsigned n = 0; n < WXSIZEOF(equivAttrs); n++ )
 453     {
 454         const EquivAttr& ea = equivAttrs[n];
 455         if ( styleParams.HasParam(ea.style) && !HasParam(ea.attr) )
 456         {
 457             m_ParamNames.Add(ea.attr);
 458             m_ParamValues.Add(styleParams.GetParam(ea.style));
 459         }
 460     }
 461 }
 462
 463 wxHtmlTag::~wxHtmlTag()
 464 {
 465     wxHtmlTag *t1, *t2;
 466     t1 = m_FirstChild;
 467     while (t1)
 468     {
 469         t2 = t1->GetNextSibling();
 470         delete t1;
 471         t1 = t2;
 472     }
 473 }
 474
 475 bool wxHtmlTag::HasParam(const wxString& par) const
 476 {
 477     return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
 478 }
 479
 480 wxString wxHtmlTag::GetParam(const wxString& par, bool with_quotes) const
 481 {
 482     int index = m_ParamNames.Index(par, false);
 483     if (index == wxNOT_FOUND)
 484         return wxGetEmptyString();
 485     if (with_quotes)
 486     {
 487         // VS: backward compatibility, seems to be never used by wxHTML...
 488         wxString s;
 489         s << wxT('"') << m_ParamValues[index] << wxT('"');
 490         return s;
 491     }
 492     else
 493         return m_ParamValues[index];
 494 }
 495
 496 int wxHtmlTag::ScanParam(const wxString& par,
 497                          const char *format,
 498                          void *param) const
 499 {
 500     wxString parval = GetParam(par);
 501     return wxSscanf(parval, format, param);
 502 }
 503
 504 int wxHtmlTag::ScanParam(const wxString& par,
 505                          const wchar_t *format,
 506                          void *param) const
 507 {
 508     wxString parval = GetParam(par);
 509     return wxSscanf(parval, format, param);
 510 }
 511
 512 /* static */
 513 bool wxHtmlTag::ParseAsColour(const wxString& str, wxColour *clr)
 514 {
 515     wxCHECK_MSG( clr, false, wxT("invalid colour argument") );
 516
 517     // handle colours defined in HTML 4.0 first:
 518     if (str.length() > 1 && str[0] != wxT('#'))
 519     {
 520         #define HTML_COLOUR(name, r, g, b)              \
 521             if (str.IsSameAs(wxS(name), false))         \
 522                 { clr->Set(r, g, b); return true; }
 523         HTML_COLOUR("black",   0x00,0x00,0x00)
 524         HTML_COLOUR("silver",  0xC0,0xC0,0xC0)
 525         HTML_COLOUR("gray",    0x80,0x80,0x80)
 526         HTML_COLOUR("white",   0xFF,0xFF,0xFF)
 527         HTML_COLOUR("maroon",  0x80,0x00,0x00)
 528         HTML_COLOUR("red",     0xFF,0x00,0x00)
 529         HTML_COLOUR("purple",  0x80,0x00,0x80)
 530         HTML_COLOUR("fuchsia", 0xFF,0x00,0xFF)
 531         HTML_COLOUR("green",   0x00,0x80,0x00)
 532         HTML_COLOUR("lime",    0x00,0xFF,0x00)
 533         HTML_COLOUR("olive",   0x80,0x80,0x00)
 534         HTML_COLOUR("yellow",  0xFF,0xFF,0x00)
 535         HTML_COLOUR("navy",    0x00,0x00,0x80)
 536         HTML_COLOUR("blue",    0x00,0x00,0xFF)
 537         HTML_COLOUR("teal",    0x00,0x80,0x80)
 538         HTML_COLOUR("aqua",    0x00,0xFF,0xFF)
 539         #undef HTML_COLOUR
 540     }
 541
 542     // then try to parse #rrggbb representations or set from other well
 543     // known names (note that this doesn't strictly conform to HTML spec,
 544     // but it doesn't do real harm -- but it *must* be done after the standard
 545     // colors are handled above):
 546     if (clr->Set(str))
 547         return true;
 548
 549     return false;
 550 }
 551
 552 bool wxHtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
 553 {
 554     const wxString str = GetParam(par);
 555     return !str.empty() && ParseAsColour(str, clr);
 556 }
 557
 558 bool wxHtmlTag::GetParamAsInt(const wxString& par, int *clr) const
 559 {
 560     if ( !HasParam(par) )
 561         return false;
 562
 563     long i;
 564     if ( !GetParam(par).ToLong(&i) )
 565         return false;
 566
 567     *clr = (int)i;
 568     return true;
 569 }
 570
 571 wxString wxHtmlTag::GetAllParams() const
 572 {
 573     // VS: this function is for backward compatibility only,
 574     //     never used by wxHTML
 575     wxString s;
 576     size_t cnt = m_ParamNames.GetCount();
 577     for (size_t i = 0; i < cnt; i++)
 578     {
 579         s << m_ParamNames[i];
 580         s << wxT('=');
 581         if (m_ParamValues[i].Find(wxT('"')) != wxNOT_FOUND)
 582             s << wxT('\'') << m_ParamValues[i] << wxT('\'');
 583         else
 584             s << wxT('"') << m_ParamValues[i] << wxT('"');
 585     }
 586     return s;
 587 }
 588
 589 wxHtmlTag *wxHtmlTag::GetFirstSibling() const
 590 {
 591     if (m_Parent)
 592         return m_Parent->m_FirstChild;
 593     else
 594     {
 595         wxHtmlTag *cur = (wxHtmlTag*)this;
 596         while (cur->m_Prev)
 597             cur = cur->m_Prev;
 598         return cur;
 599     }
 600 }
 601
 602 wxHtmlTag *wxHtmlTag::GetLastSibling() const
 603 {
 604     if (m_Parent)
 605         return m_Parent->m_LastChild;
 606     else
 607     {
 608         wxHtmlTag *cur = (wxHtmlTag*)this;
 609         while (cur->m_Next)
 610             cur = cur->m_Next;
 611         return cur;
 612     }
 613 }
 614
 615 wxHtmlTag *wxHtmlTag::GetNextTag() const
 616 {
 617     if (m_FirstChild) return m_FirstChild;
 618     if (m_Next) return m_Next;
 619     wxHtmlTag *cur = m_Parent;
 620     if (!cur) return NULL;
 621     while (cur->m_Parent && !cur->m_Next)
 622         cur = cur->m_Parent;
 623     return cur->m_Next;
 624 }
 625
 626 #endif