src/html/htmltag.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        src/html/htmltag.cpp
   3 // Purpose:     wxHtmlTag class (represents single tag)
   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 #if wxUSE_HTML
  16
  17 #include "wx/html/htmltag.h"
  18
  19 #ifndef WX_PRECOMP
  20     #include "wx/colour.h"
  21     #include "wx/wxcrtvararg.h"
  22 #endif
  23
  24 #include "wx/html/htmlpars.h"
  25 #include "wx/html/styleparams.h"
  26
  27 #include "wx/vector.h"
  28
  29 #include <stdio.h> // for vsscanf
  30 #include <stdarg.h>
  31
  32 //-----------------------------------------------------------------------------
  33 // wxHtmlTagsCache
  34 //-----------------------------------------------------------------------------
  35
  36 struct wxHtmlCacheItem
  37 {
  38     // this is "pos" value passed to wxHtmlTag's constructor.
  39     // it is position of '<' character of the tag
  40     wxString::const_iterator Key;
  41
  42     // Tag type
  43     enum Type
  44     {
  45         Type_Normal, // normal tag with a matching ending tag
  46         Type_NoMatchingEndingTag, // there's no ending tag for this tag
  47         Type_EndingTag // this is ending tag </..>
  48     };
  49     Type type;
  50
  51     // end positions for the tag:
  52     // end1 is '<' of ending tag,
  53     // end2 is '>' or both are
  54     wxString::const_iterator End1, End2;
  55
  56     // name of this tag
  57     wxChar *Name;
  58 };
  59
  60 // NB: this is an empty class and not typedef because of forward declaration
  61 class wxHtmlTagsCacheData : public wxVector<wxHtmlCacheItem>
  62 {
  63 };
  64
  65 bool wxIsCDATAElement(const wxChar *tag)
  66 {
  67     return (wxStrcmp(tag, wxT("SCRIPT")) == 0) ||
  68            (wxStrcmp(tag, wxT("STYLE")) == 0);
  69 }
  70
  71 bool wxIsCDATAElement(const wxString& tag)
  72 {
  73     return (wxStrcmp(tag.wx_str(), wxS("SCRIPT")) == 0) ||
  74            (wxStrcmp(tag.wx_str(), wxS("STYLE")) == 0);
  75 }
  76
  77 wxHtmlTagsCache::wxHtmlTagsCache(const wxString& source)
  78 {
  79     m_Cache = new wxHtmlTagsCacheData;
  80     m_CachePos = 0;
  81
  82     wxChar tagBuffer[256];
  83
  84     const wxString::const_iterator end = source.end();
  85     for ( wxString::const_iterator pos = source.begin(); pos < end; ++pos )
  86     {
  87         if (*pos != wxT('<'))
  88             continue;
  89
  90         // possible tag start found:
  91
  92         // don't cache comment tags
  93         if ( wxHtmlParser::SkipCommentTag(pos, end) )
  94             continue;
  95
  96         // Remember the starting tag position.
  97         wxString::const_iterator stpos = pos++;
  98
  99         // And look for the ending one.
 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         while (pos < end && *pos != wxT('>'))
 111             ++pos;
 112
 113         if ( pos == end )
 114         {
 115             // We didn't find a closing bracket, this is not a valid tag after
 116             // all. Notice that we need to roll back pos to avoid creating an
 117             // invalid iterator when "++pos" is done in the loop statement.
 118             --pos;
 119
 120             continue;
 121         }
 122
 123         // We have a valid tag, add it to the cache.
 124         size_t tg = Cache().size();
 125         Cache().push_back(wxHtmlCacheItem());
 126         Cache()[tg].Key = stpos;
 127         Cache()[tg].Name = new wxChar[i+1];
 128         memcpy(Cache()[tg].Name, tagBuffer, (i+1)*sizeof(wxChar));
 129
 130         if ((stpos+1) < end && *(stpos+1) == wxT('/')) // ending tag:
 131         {
 132             Cache()[tg].type = wxHtmlCacheItem::Type_EndingTag;
 133             // find matching begin tag:
 134             for (i = tg; i >= 0; i--)
 135             {
 136                 if ((Cache()[i].type == wxHtmlCacheItem::Type_NoMatchingEndingTag) && (wxStrcmp(Cache()[i].Name, tagBuffer+1) == 0))
 137                 {
 138                     Cache()[i].type = wxHtmlCacheItem::Type_Normal;
 139                     Cache()[i].End1 = stpos;
 140                     Cache()[i].End2 = pos + 1;
 141                     break;
 142                 }
 143             }
 144         }
 145         else
 146         {
 147             Cache()[tg].type = wxHtmlCacheItem::Type_NoMatchingEndingTag;
 148
 149             if (wxIsCDATAElement(tagBuffer))
 150             {
 151                 // store the orig pos in case we are missing the closing
 152                 // tag (see below)
 153                 const wxString::const_iterator old_pos = pos;
 154                 bool foundCloseTag = false;
 155
 156                 // find next matching tag
 157                 int tag_len = wxStrlen(tagBuffer);
 158                 while (pos < end)
 159                 {
 160                     // find the ending tag
 161                     while (pos + 1 < end &&
 162                            (*pos != '<' || *(pos+1) != '/'))
 163                         ++pos;
 164                     if (*pos == '<')
 165                         ++pos;
 166
 167                     // see if it matches
 168                     int match_pos = 0;
 169                     while (pos < end && match_pos < tag_len )
 170                     {
 171                         wxChar c = *pos;
 172                         if ( c == '>' || c == '<' )
 173                             break;
 174
 175                         // cast to wxChar needed to suppress warning in
 176                         // Unicode build
 177                         if ((wxChar)wxToupper(c) == tagBuffer[match_pos])
 178                         {
 179                             ++match_pos;
 180                         }
 181                         else if (c == wxT(' ') || c == wxT('\n') ||
 182                             c == wxT('\r') || c == wxT('\t'))
 183                         {
 184                             // need to skip over these
 185                         }
 186                         else
 187                         {
 188                             match_pos = 0;
 189                         }
 190                         ++pos;
 191                     }
 192
 193                     // found a match
 194                     if (match_pos == tag_len)
 195                     {
 196                         pos = pos - tag_len - 3;
 197                         foundCloseTag = true;
 198                         break;
 199                     }
 200                     else // keep looking for the closing tag
 201                     {
 202                         ++pos;
 203                     }
 204                 }
 205                 if (!foundCloseTag)
 206                 {
 207                     // we didn't find closing tag; this means the markup
 208                     // is incorrect and the best thing we can do is to
 209                     // ignore the unclosed tag and continue parsing as if
 210                     // it didn't exist:
 211                     pos = old_pos;
 212                 }
 213             }
 214         }
 215     }
 216
 217     // ok, we're done, now we'll free .Name members of cache - we don't need it anymore:
 218     for ( wxHtmlTagsCacheData::iterator i = Cache().begin();
 219           i != Cache().end(); ++i )
 220     {
 221         wxDELETEA(i->Name);
 222     }
 223 }
 224
 225 wxHtmlTagsCache::~wxHtmlTagsCache()
 226 {
 227     delete m_Cache;
 228 }
 229
 230 void wxHtmlTagsCache::QueryTag(const wxString::const_iterator& at,
 231                                const wxString::const_iterator& inputEnd,
 232                                wxString::const_iterator *end1,
 233                                wxString::const_iterator *end2,
 234                                bool *hasEnding)
 235 {
 236     if (Cache().empty())
 237     {
 238         *end1 =
 239         *end2 = inputEnd;
 240         *hasEnding = true;
 241         return;
 242     }
 243
 244     if (Cache()[m_CachePos].Key != at)
 245     {
 246         int delta = (at < Cache()[m_CachePos].Key) ? -1 : 1;
 247         do
 248         {
 249             m_CachePos += delta;
 250
 251             if ( m_CachePos < 0 || m_CachePos >= (int)Cache().size() )
 252             {
 253                 if ( m_CachePos < 0 )
 254                     m_CachePos = 0;
 255                 else
 256                     m_CachePos = Cache().size() - 1;
 257                 // something is very wrong with HTML, give up by returning an
 258                 // impossibly large value which is going to be ignored by the
 259                 // caller
 260                 *end1 =
 261                 *end2 = inputEnd;
 262                 *hasEnding = true;
 263                 return;
 264             }
 265         }
 266         while (Cache()[m_CachePos].Key != at);
 267     }
 268
 269     switch ( Cache()[m_CachePos].type )
 270     {
 271         case wxHtmlCacheItem::Type_Normal:
 272             *end1 = Cache()[m_CachePos].End1;
 273             *end2 = Cache()[m_CachePos].End2;
 274             *hasEnding = true;
 275             break;
 276
 277         case wxHtmlCacheItem::Type_EndingTag:
 278             wxFAIL_MSG("QueryTag called for ending tag - can't be");
 279             // but if it does happen, fall through, better than crashing
 280
 281         case wxHtmlCacheItem::Type_NoMatchingEndingTag:
 282             // If input HTML is invalid and there's no closing tag for this
 283             // one, pretend that it runs all the way to the end of input
 284             *end1 = inputEnd;
 285             *end2 = inputEnd;
 286             *hasEnding = false;
 287             break;
 288     }
 289 }
 290
 291
 292
 293
 294 //-----------------------------------------------------------------------------
 295 // wxHtmlTag
 296 //-----------------------------------------------------------------------------
 297
 298 wxHtmlTag::wxHtmlTag(wxHtmlTag *parent,
 299                      const wxString *source,
 300                      const wxString::const_iterator& pos,
 301                      const wxString::const_iterator& end_pos,
 302                      wxHtmlTagsCache *cache,
 303                      wxHtmlEntitiesParser *entParser)
 304 {
 305     /* Setup DOM relations */
 306
 307     m_Next = NULL;
 308     m_FirstChild = m_LastChild = NULL;
 309     m_Parent = parent;
 310     if (parent)
 311     {
 312         m_Prev = m_Parent->m_LastChild;
 313         if (m_Prev == NULL)
 314             m_Parent->m_FirstChild = this;
 315         else
 316             m_Prev->m_Next = this;
 317         m_Parent->m_LastChild = this;
 318     }
 319     else
 320         m_Prev = NULL;
 321
 322     /* Find parameters and their values: */
 323
 324     wxChar c wxDUMMY_INITIALIZE(0);
 325
 326     // fill-in name, params and begin pos:
 327     wxString::const_iterator i(pos+1);
 328
 329     // find tag's name and convert it to uppercase:
 330     while ((i < end_pos) &&
 331            ((c = *(i++)) != wxT(' ') && c != wxT('\r') &&
 332              c != wxT('\n') && c != wxT('\t') &&
 333              c != wxT('>') && c != wxT('/')))
 334     {
 335         if ((c >= wxT('a')) && (c <= wxT('z')))
 336             c -= (wxT('a') - wxT('A'));
 337         m_Name << c;
 338     }
 339
 340     // if the tag has parameters, read them and "normalize" them,
 341     // i.e. convert to uppercase, replace whitespaces by spaces and
 342     // remove whitespaces around '=':
 343     if (*(i-1) != wxT('>'))
 344     {
 345         #define IS_WHITE(c) (c == wxT(' ') || c == wxT('\r') || \
 346                              c == wxT('\n') || c == wxT('\t'))
 347         wxString pname, pvalue;
 348         wxChar quote;
 349         enum
 350         {
 351             ST_BEFORE_NAME = 1,
 352             ST_NAME,
 353             ST_BEFORE_EQ,
 354             ST_BEFORE_VALUE,
 355             ST_VALUE
 356         } state;
 357
 358         quote = 0;
 359         state = ST_BEFORE_NAME;
 360         while (i < end_pos)
 361         {
 362             c = *(i++);
 363
 364             if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
 365             {
 366                 if (state == ST_BEFORE_EQ || state == ST_NAME)
 367                 {
 368                     m_ParamNames.Add(pname);
 369                     m_ParamValues.Add(wxGetEmptyString());
 370                 }
 371                 else if (state == ST_VALUE && quote == 0)
 372                 {
 373                     m_ParamNames.Add(pname);
 374                     if (entParser)
 375                         m_ParamValues.Add(entParser->Parse(pvalue));
 376                     else
 377                         m_ParamValues.Add(pvalue);
 378                 }
 379                 break;
 380             }
 381             switch (state)
 382             {
 383                 case ST_BEFORE_NAME:
 384                     if (!IS_WHITE(c))
 385                     {
 386                         pname = c;
 387                         state = ST_NAME;
 388                     }
 389                     break;
 390                 case ST_NAME:
 391                     if (IS_WHITE(c))
 392                         state = ST_BEFORE_EQ;
 393                     else if (c == wxT('='))
 394                         state = ST_BEFORE_VALUE;
 395                     else
 396                         pname << c;
 397                     break;
 398                 case ST_BEFORE_EQ:
 399                     if (c == wxT('='))
 400                         state = ST_BEFORE_VALUE;
 401                     else if (!IS_WHITE(c))
 402                     {
 403                         m_ParamNames.Add(pname);
 404                         m_ParamValues.Add(wxGetEmptyString());
 405                         pname = c;
 406                         state = ST_NAME;
 407                     }
 408                     break;
 409                 case ST_BEFORE_VALUE:
 410                     if (!IS_WHITE(c))
 411                     {
 412                         if (c == wxT('"') || c == wxT('\''))
 413                             quote = c, pvalue = wxGetEmptyString();
 414                         else
 415                             quote = 0, pvalue = c;
 416                         state = ST_VALUE;
 417                     }
 418                     break;
 419                 case ST_VALUE:
 420                     if ((quote != 0 && c == quote) ||
 421                         (quote == 0 && IS_WHITE(c)))
 422                     {
 423                         m_ParamNames.Add(pname);
 424                         if (quote == 0)
 425                         {
 426                             // VS: backward compatibility, no real reason,
 427                             //     but wxHTML code relies on this... :(
 428                             pvalue.MakeUpper();
 429                         }
 430                         if (entParser)
 431                             m_ParamValues.Add(entParser->Parse(pvalue));
 432                         else
 433                             m_ParamValues.Add(pvalue);
 434                         state = ST_BEFORE_NAME;
 435                     }
 436                     else
 437                         pvalue << c;
 438                     break;
 439             }
 440         }
 441
 442         #undef IS_WHITE
 443     }
 444     m_Begin = i;
 445     cache->QueryTag(pos, source->end(), &m_End1, &m_End2, &m_hasEnding);
 446     if (m_End1 > end_pos) m_End1 = end_pos;
 447     if (m_End2 > end_pos) m_End2 = end_pos;
 448
 449 #if WXWIN_COMPATIBILITY_2_8
 450     m_sourceStart = source->begin();
 451 #endif
 452
 453     // Try to parse any style parameters that can be handled simply by
 454     // converting them to the equivalent HTML 3 attributes: this is a far cry
 455     // from perfect but better than nothing.
 456     static const struct EquivAttr
 457     {
 458         const char *style;
 459         const char *attr;
 460     } equivAttrs[] =
 461     {
 462         { "text-align",         "ALIGN"         },
 463         { "width",              "WIDTH"         },
 464         { "vertical-align",     "VALIGN"        },
 465         { "background",         "BGCOLOR"       },
 466         { "background-color",   "BGCOLOR"       },
 467     };
 468
 469     wxHtmlStyleParams styleParams(*this);
 470     for ( unsigned n = 0; n < WXSIZEOF(equivAttrs); n++ )
 471     {
 472         const EquivAttr& ea = equivAttrs[n];
 473         if ( styleParams.HasParam(ea.style) && !HasParam(ea.attr) )
 474         {
 475             m_ParamNames.Add(ea.attr);
 476             m_ParamValues.Add(styleParams.GetParam(ea.style));
 477         }
 478     }
 479 }
 480
 481 wxHtmlTag::~wxHtmlTag()
 482 {
 483     wxHtmlTag *t1, *t2;
 484     t1 = m_FirstChild;
 485     while (t1)
 486     {
 487         t2 = t1->GetNextSibling();
 488         delete t1;
 489         t1 = t2;
 490     }
 491 }
 492
 493 bool wxHtmlTag::HasParam(const wxString& par) const
 494 {
 495     return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
 496 }
 497
 498 wxString wxHtmlTag::GetParam(const wxString& par, bool with_quotes) const
 499 {
 500     int index = m_ParamNames.Index(par, false);
 501     if (index == wxNOT_FOUND)
 502         return wxGetEmptyString();
 503     if (with_quotes)
 504     {
 505         // VS: backward compatibility, seems to be never used by wxHTML...
 506         wxString s;
 507         s << wxT('"') << m_ParamValues[index] << wxT('"');
 508         return s;
 509     }
 510     else
 511         return m_ParamValues[index];
 512 }
 513
 514 int wxHtmlTag::ScanParam(const wxString& par,
 515                          const char *format,
 516                          void *param) const
 517 {
 518     wxString parval = GetParam(par);
 519     return wxSscanf(parval, format, param);
 520 }
 521
 522 int wxHtmlTag::ScanParam(const wxString& par,
 523                          const wchar_t *format,
 524                          void *param) const
 525 {
 526     wxString parval = GetParam(par);
 527     return wxSscanf(parval, format, param);
 528 }
 529
 530 /* static */
 531 bool wxHtmlTag::ParseAsColour(const wxString& str, wxColour *clr)
 532 {
 533     wxCHECK_MSG( clr, false, wxT("invalid colour argument") );
 534
 535     // handle colours defined in HTML 4.0 first:
 536     if (str.length() > 1 && str[0] != wxT('#'))
 537     {
 538         #define HTML_COLOUR(name, r, g, b)              \
 539             if (str.IsSameAs(wxS(name), false))         \
 540                 { clr->Set(r, g, b); return true; }
 541         HTML_COLOUR("black",   0x00,0x00,0x00)
 542         HTML_COLOUR("silver",  0xC0,0xC0,0xC0)
 543         HTML_COLOUR("gray",    0x80,0x80,0x80)
 544         HTML_COLOUR("white",   0xFF,0xFF,0xFF)
 545         HTML_COLOUR("maroon",  0x80,0x00,0x00)
 546         HTML_COLOUR("red",     0xFF,0x00,0x00)
 547         HTML_COLOUR("purple",  0x80,0x00,0x80)
 548         HTML_COLOUR("fuchsia", 0xFF,0x00,0xFF)
 549         HTML_COLOUR("green",   0x00,0x80,0x00)
 550         HTML_COLOUR("lime",    0x00,0xFF,0x00)
 551         HTML_COLOUR("olive",   0x80,0x80,0x00)
 552         HTML_COLOUR("yellow",  0xFF,0xFF,0x00)
 553         HTML_COLOUR("navy",    0x00,0x00,0x80)
 554         HTML_COLOUR("blue",    0x00,0x00,0xFF)
 555         HTML_COLOUR("teal",    0x00,0x80,0x80)
 556         HTML_COLOUR("aqua",    0x00,0xFF,0xFF)
 557         #undef HTML_COLOUR
 558     }
 559
 560     // then try to parse #rrggbb representations or set from other well
 561     // known names (note that this doesn't strictly conform to HTML spec,
 562     // but it doesn't do real harm -- but it *must* be done after the standard
 563     // colors are handled above):
 564     if (clr->Set(str))
 565         return true;
 566
 567     return false;
 568 }
 569
 570 bool wxHtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
 571 {
 572     const wxString str = GetParam(par);
 573     return !str.empty() && ParseAsColour(str, clr);
 574 }
 575
 576 bool wxHtmlTag::GetParamAsInt(const wxString& par, int *clr) const
 577 {
 578     if ( !HasParam(par) )
 579         return false;
 580
 581     long i;
 582     if ( !GetParam(par).ToLong(&i) )
 583         return false;
 584
 585     *clr = (int)i;
 586     return true;
 587 }
 588
 589 bool
 590 wxHtmlTag::GetParamAsIntOrPercent(const wxString& par,
 591                                   int* value,
 592                                   bool& isPercent) const
 593 {
 594     const wxString param = GetParam(par);
 595     if ( param.empty() )
 596         return false;
 597
 598     wxString num;
 599     if ( param.EndsWith("%", &num) )
 600     {
 601         isPercent = true;
 602     }
 603     else
 604     {
 605         isPercent = false;
 606         num = param;
 607     }
 608
 609     long lValue;
 610     if ( !num.ToLong(&lValue) )
 611         return false;
 612
 613     if ( lValue > INT_MAX || lValue < INT_MIN )
 614         return false;
 615
 616     *value = static_cast<int>(lValue);
 617
 618     return true;
 619 }
 620
 621 wxString wxHtmlTag::GetAllParams() const
 622 {
 623     // VS: this function is for backward compatibility only,
 624     //     never used by wxHTML
 625     wxString s;
 626     size_t cnt = m_ParamNames.GetCount();
 627     for (size_t i = 0; i < cnt; i++)
 628     {
 629         s << m_ParamNames[i];
 630         s << wxT('=');
 631         if (m_ParamValues[i].Find(wxT('"')) != wxNOT_FOUND)
 632             s << wxT('\'') << m_ParamValues[i] << wxT('\'');
 633         else
 634             s << wxT('"') << m_ParamValues[i] << wxT('"');
 635     }
 636     return s;
 637 }
 638
 639 wxHtmlTag *wxHtmlTag::GetFirstSibling() const
 640 {
 641     if (m_Parent)
 642         return m_Parent->m_FirstChild;
 643     else
 644     {
 645         wxHtmlTag *cur = (wxHtmlTag*)this;
 646         while (cur->m_Prev)
 647             cur = cur->m_Prev;
 648         return cur;
 649     }
 650 }
 651
 652 wxHtmlTag *wxHtmlTag::GetLastSibling() const
 653 {
 654     if (m_Parent)
 655         return m_Parent->m_LastChild;
 656     else
 657     {
 658         wxHtmlTag *cur = (wxHtmlTag*)this;
 659         while (cur->m_Next)
 660             cur = cur->m_Next;
 661         return cur;
 662     }
 663 }
 664
 665 wxHtmlTag *wxHtmlTag::GetNextTag() const
 666 {
 667     if (m_FirstChild) return m_FirstChild;
 668     if (m_Next) return m_Next;
 669     wxHtmlTag *cur = m_Parent;
 670     if (!cur) return NULL;
 671     while (cur->m_Parent && !cur->m_Next)
 672         cur = cur->m_Parent;
 673     return cur->m_Next;
 674 }
 675
 676 #endif