xbmc/utils/HTMLUtil.cpp

   1 /*
   2  *      Copyright (C) 2005-2008 Team XBMC
   3  *      http://www.xbmc.org
   4  *
   5  *  This Program is free software; you can redistribute it and/or modify
   6  *  it under the terms of the GNU General Public License as published by
   7  *  the Free Software Foundation; either version 2, or (at your option)
   8  *  any later version.
   9  *
  10  *  This Program is distributed in the hope that it will be useful,
  11  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  *  GNU General Public License for more details.
  14  *
  15  *  You should have received a copy of the GNU General Public License
  16  *  along with XBMC; see the file COPYING.  If not, write to
  17  *  the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
  18  *  http://www.gnu.org/copyleft/gpl.html
  19  *
  20  */
  21
  22 #include "CharsetConverter.h"
  23 #include "HTMLUtil.h"
  24
  25 using namespace std;
  26 using namespace HTML;
  27
  28
  29 CHTMLUtil::CHTMLUtil(void)
  30 {}
  31
  32 CHTMLUtil::~CHTMLUtil(void)
  33 {}
  34
  35 int CHTMLUtil::FindTag(const CStdString& strHTML, const CStdString& strTag, CStdString& strtagFound, int iPos) const
  36 {
  37   CStdString strHTMLLow = strHTML;
  38   CStdString strTagLow = strTag;
  39   strHTMLLow.ToLower();
  40   strTagLow.ToLower();
  41   strtagFound = "";
  42   int iStart = strHTMLLow.Find(strTag, iPos);
  43   if (iStart < 0) return -1;
  44   int iEnd = strHTMLLow.Find(">", iStart);
  45   if (iEnd < 0) iEnd = (int)strHTMLLow.size();
  46   strtagFound = strHTMLLow.Mid(iStart, (iEnd + 1) - iStart);
  47   return iStart;
  48 }
  49
  50 int CHTMLUtil::FindClosingTag(const CStdString& strHTML, const CStdString& strTag, CStdString& strtagFound, int iPos) const
  51 {
  52   CStdString strHTMLLow = strHTML;
  53   CStdString strTagLow = strTag;
  54   strHTMLLow.ToLower();
  55   strTagLow.ToLower();
  56   strtagFound = "";
  57   int iStart = strHTMLLow.Find("</" + strTag, iPos);
  58   if (iStart < 0) return -1;
  59   int iOpenStart = strHTMLLow.Find("<" + strTag, iPos);
  60   while (iOpenStart < iStart && iOpenStart != -1)
  61   {
  62     iStart = strHTMLLow.Find("</" + strTag, iStart + 1);
  63     iOpenStart = strHTMLLow.Find("<" + strTag, iOpenStart + 1);
  64   }
  65
  66   int iEnd = strHTMLLow.Find(">", iStart);
  67   if (iEnd < 0) iEnd = (int)strHTMLLow.size();
  68   strtagFound = strHTMLLow.Mid(iStart, (iEnd + 1) - iStart);
  69   return iStart;
  70 }
  71
  72 void CHTMLUtil::getValueOfTag(const CStdString& strTagAndValue, CStdString& strValue)
  73 {
  74   // strTagAndValue contains:
  75   // like <a href=blablabla.....>value</a>
  76   strValue = strTagAndValue;
  77   int iStart = strTagAndValue.Find(">");
  78   int iEnd = strTagAndValue.Find("<", iStart + 1);
  79   if (iStart >= 0 && iEnd >= 0)
  80   {
  81     iStart++;
  82     strValue = strTagAndValue.Mid(iStart, iEnd - iStart);
  83   }
  84 }
  85
  86 void CHTMLUtil::getAttributeOfTag(const CStdString& strTagAndValue, const CStdString& strTag, CStdString& strValue)
  87 {
  88   // strTagAndValue contains:
  89   // like <a href=""value".....
  90   strValue = strTagAndValue;
  91   int iStart = strTagAndValue.Find(strTag);
  92   if (iStart < 0) return ;
  93   iStart += (int)strTag.size();
  94   while (strTagAndValue[iStart + 1] == 0x20 || strTagAndValue[iStart + 1] == 0x27 || strTagAndValue[iStart + 1] == 34) iStart++;
  95   int iEnd = iStart + 1;
  96   while (strTagAndValue[iEnd] != 0x27 && strTagAndValue[iEnd] != 0x20 && strTagAndValue[iEnd] != 34 && strTagAndValue[iEnd] != '>') iEnd++;
  97   if (iStart >= 0 && iEnd >= 0)
  98   {
  99     strValue = strTagAndValue.Mid(iStart, iEnd - iStart);
 100   }
 101 }
 102
 103 void CHTMLUtil::RemoveTags(CStdString& strHTML)
 104 {
 105   int iNested = 0;
 106   CStdString strReturn = "";
 107   for (int i = 0; i < (int) strHTML.size(); ++i)
 108   {
 109     if (strHTML[i] == '<') iNested++;
 110     else if (strHTML[i] == '>') iNested--;
 111     else
 112     {
 113       if (!iNested)
 114       {
 115         strReturn += strHTML[i];
 116       }
 117     }
 118   }
 119
 120   strHTML = strReturn;
 121 }
 122
 123 typedef struct
 124 {
 125   const wchar_t* html;
 126   const wchar_t  w;
 127 } HTMLMapping;
 128
 129 static const HTMLMapping mappings[] =
 130   {{L"&amp;",     0x0026},
 131    {L"&apos;",    0x0027},
 132    {L"&acute;",   0x00B4},
 133    {L"&agrave;",  0x00E0},
 134    {L"&aacute;",  0x00E1},
 135    {L"&acirc;",   0x00E2},
 136    {L"&atilde;",  0x00E3},
 137    {L"&auml;",    0x00E4},
 138    {L"&aring;",   0x00E5},
 139    {L"&aelig;",   0x00E6},
 140    {L"&Agrave;",  0x00C0},
 141    {L"&Aacute;",  0x00C1},
 142    {L"&Acirc;",   0x00C2},
 143    {L"&Atilde;",  0x00C3},
 144    {L"&Auml;",    0x00C4},
 145    {L"&Aring;",   0x00C5},
 146    {L"&AElig;",   0x00C6},
 147    {L"&bdquo;",   0x201E},
 148    {L"&brvbar;",  0x00A6},
 149    {L"&bull;",    0x2022},
 150    {L"&bullet;",  0x2022},
 151    {L"&cent;",    0x00A2},
 152    {L"&circ;",    0x02C6},
 153    {L"&curren;",  0x00A4},
 154    {L"&copy;",    0x00A9},
 155    {L"&cedil;",   0x00B8},
 156    {L"&Ccedil;",  0x00C7},
 157    {L"&ccedil;",  0x00E7},
 158    {L"&dagger;",  0x2020},
 159    {L"&deg;",     0x00B0},
 160    {L"&divide;",  0x00F7},
 161    {L"&Dagger;",  0x2021},
 162    {L"&egrave;",  0x00E8},
 163    {L"&eacute;",  0x00E9},
 164    {L"&ecirc;",   0x00EA},
 165    {L"&emsp;",    0x2003},
 166    {L"&ensp;",    0x2002},
 167    {L"&euml;",    0x00EB},
 168    {L"&eth;",     0x00F0},
 169    {L"&euro;",    0x20AC},
 170    {L"&Egrave;",  0x00C8},
 171    {L"&Eacute;",  0x00C9},
 172    {L"&Ecirc;",   0x00CA},
 173    {L"&Euml;",    0x00CB},
 174    {L"&ETH;",     0x00D0},
 175    {L"&quot;",    0x0022},
 176    {L"&frasl;",   0x2044},
 177    {L"&frac14;",  0x00BC},
 178    {L"&frac12;",  0x00BD},
 179    {L"&frac34;",  0x00BE},
 180    {L"&gt;",      0x003E},
 181    {L"&hellip;",  0x2026},
 182    {L"&iexcl;",   0x00A1},
 183    {L"&iquest;",  0x00BF},
 184    {L"&igrave;",  0x00EC},
 185    {L"&iacute;",  0x00ED},
 186    {L"&icirc;",   0x00EE},
 187    {L"&iuml;",    0x00EF},
 188    {L"&Igrave;",  0x00CC},
 189    {L"&Iacute;",  0x00CD},
 190    {L"&Icirc;",   0x00CE},
 191    {L"&Iuml;",    0x00CF},
 192    {L"&lrm;",     0x200E},
 193    {L"&lt;",      0x003C},
 194    {L"&laquo;",   0x00AB},
 195    {L"&ldquo;",   0x201C},
 196    {L"&lsaquo;",  0x2039},
 197    {L"&lsquo;",   0x2018},
 198    {L"&macr;",    0x00AF},
 199    {L"&micro;",   0x00B5},
 200    {L"&middot;",  0x00B7},
 201    {L"&mdash;",   0x2014},
 202    {L"&nbsp;",    0x00A0},
 203    {L"&ndash;",   0x2013},
 204    {L"&ntilde;",  0x00F1},
 205    {L"&not;",     0x00AC},
 206    {L"&Ntilde;",  0x00D1},
 207    {L"&ordf;",    0x00AA},
 208    {L"&ordm;",    0x00BA},
 209    {L"&oelig;",   0x0153},
 210    {L"&ograve;",  0x00F2},
 211    {L"&oacute;",  0x00F3},
 212    {L"&ocirc;",   0x00F4},
 213    {L"&otilde;",  0x00F5},
 214    {L"&ouml;",    0x00F6},
 215    {L"&oslash;",  0x00F8},
 216    {L"&OElig;",   0x0152},
 217    {L"&Ograve;",  0x00D2},
 218    {L"&Oacute;",  0x00D3},
 219    {L"&Ocirc;",   0x00D4},
 220    {L"&Otilde;",  0x00D5},
 221    {L"&Ouml;",    0x00D6},
 222    {L"&Oslash;",  0x00D8},
 223    {L"&para;",    0x00B6},
 224    {L"&permil;",  0x2030},
 225    {L"&plusmn;",  0x00B1},
 226    {L"&pound;",   0x00A3},
 227    {L"&raquo;",   0x00BB},
 228    {L"&rdquo;",   0x201D},
 229    {L"&reg;",     0x00AE},
 230    {L"&rlm;",     0x200F},
 231    {L"&rsaquo;",  0x203A},
 232    {L"&rsquo;",   0x2019},
 233    {L"&sbquo;",   0x201A},
 234    {L"&scaron;",  0x0161},
 235    {L"&sect;",    0x00A7},
 236    {L"&shy;",     0x00AD},
 237    {L"&sup1;",    0x00B9},
 238    {L"&sup2;",    0x00B2},
 239    {L"&sup3;",    0x00B3},
 240    {L"&szlig;",   0x00DF},
 241    {L"&Scaron;",  0x0160},
 242    {L"&thinsp;",  0x2009},
 243    {L"&thorn;",   0x00FE},
 244    {L"&tilde;",   0x02DC},
 245    {L"&times;",   0x00D7},
 246    {L"&trade;",   0x2122},
 247    {L"&THORN;",   0x00DE},
 248    {L"&uml;",     0x00A8},
 249    {L"&ugrave;",  0x00F9},
 250    {L"&uacute;",  0x00FA},
 251    {L"&ucirc;",   0x00FB},
 252    {L"&uuml;",    0x00FC},
 253    {L"&Ugrave;",  0x00D9},
 254    {L"&Uacute;",  0x00DA},
 255    {L"&Ucirc;",   0x00DB},
 256    {L"&Uuml;",    0x00DC},
 257    {L"&yen;",     0x00A5},
 258    {L"&yuml;",    0x00FF},
 259    {L"&yacute;",  0x00FD},
 260    {L"&Yacute;",  0x00DD},
 261    {L"&Yuml;",    0x0178},
 262    {L"&zwj;",     0x200D},
 263    {L"&zwnj;",    0x200C},
 264    {NULL,         L'\0'}};
 265
 266 void CHTMLUtil::ConvertHTMLToW(const CStdStringW& strHTML, CStdStringW& strStripped)
 267 {
 268   if (strHTML.size() == 0)
 269   {
 270     strStripped.Empty();
 271     return ;
 272   }
 273   int iPos = 0;
 274   strStripped = strHTML;
 275   while (mappings[iPos].html)
 276   {
 277     strStripped.Replace(mappings[iPos].html,CStdStringW(1, mappings[iPos].w));
 278     iPos++;
 279   }
 280
 281   iPos = strStripped.Find(L"&#");
 282   while (iPos > 0 && iPos < (int)strStripped.size()-4)
 283   {
 284     int iStart = iPos + 1;
 285     iPos += 2;
 286     CStdStringW num;
 287     int base = 10;
 288     if (strStripped[iPos+1] == L'x')
 289     {
 290       base = 16;
 291       iPos++;
 292     }
 293
 294     int i=iPos;
 295     while ( iPos < (int)strStripped.size() &&
 296            (base==16?iswxdigit(strStripped[iPos]):iswdigit(strStripped[iPos])))
 297       iPos++;
 298
 299     num = strStripped.Mid(i,iPos-i);
 300     wchar_t val = (wchar_t)wcstol(num.c_str(),NULL,base);
 301     if (base == 10)
 302       num.Format(L"&#%ls;",num.c_str());
 303     else
 304       num.Format(L"&#x%ls;",num.c_str());
 305
 306     strStripped.Replace(num,CStdStringW(1,val));
 307     iPos = strStripped.Find(L"&#", iStart);
 308   }
 309 }
 310