xbmc/utils/RssReader.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 "RssReader.h"
  23 #include "utils/HTMLUtil.h"
  24 #include "../utils/Network.h"
  25 #include "Application.h"
  26 #include "CharsetConverter.h"
  27 #include "URL.h"
  28 #include "FileSystem/File.h"
  29 #include "FileSystem/FileCurl.h"
  30 #ifdef __APPLE__
  31 #include "CocoaInterface.h"
  32 #endif
  33 #include "Settings.h"
  34 #include "LocalizeStrings.h"
  35 #include "GUIRSSControl.h"
  36 #include "utils/TimeUtils.h"
  37 #include "SingleLock.h"
  38 #include "log.h"
  39
  40 using namespace std;
  41 using namespace XFILE;
  42
  43 //////////////////////////////////////////////////////////////////////
  44 // Construction/Destruction
  45 //////////////////////////////////////////////////////////////////////
  46
  47 CRssReader::CRssReader() : CThread()
  48 {
  49   m_pObserver = NULL;
  50   m_spacesBetweenFeeds = 0;
  51   m_bIsRunning = false;
  52   m_SavedScrollPos = 0;
  53 }
  54
  55 CRssReader::~CRssReader()
  56 {
  57   if (m_pObserver)
  58     m_pObserver->OnFeedRelease();
  59   StopThread();
  60   for (unsigned int i = 0; i < m_vecTimeStamps.size(); i++)
  61     delete m_vecTimeStamps[i];
  62 }
  63
  64 void CRssReader::Create(IRssObserver* aObserver, const vector<string>& aUrls, const vector<int> &times, int spacesBetweenFeeds, bool rtl)
  65 {
  66   CSingleLock lock(*this);
  67
  68   m_pObserver = aObserver;
  69   m_spacesBetweenFeeds = spacesBetweenFeeds;
  70   m_vecUrls = aUrls;
  71   m_strFeed.resize(aUrls.size());
  72   m_strColors.resize(aUrls.size());
  73   // set update times
  74   m_vecUpdateTimes = times;
  75   m_rtlText = rtl;
  76   m_requestRefresh = false;
  77
  78   // update each feed on creation
  79   for (unsigned int i=0;i<m_vecUpdateTimes.size();++i )
  80   {
  81     AddToQueue(i);
  82     SYSTEMTIME* time = new SYSTEMTIME;
  83     GetLocalTime(time);
  84     m_vecTimeStamps.push_back(time);
  85   }
  86 }
  87
  88 void CRssReader::requestRefresh()
  89 {
  90   m_requestRefresh = true;
  91 }
  92
  93 void CRssReader::AddToQueue(int iAdd)
  94 {
  95   CSingleLock lock(*this);
  96   if (iAdd < (int)m_vecUrls.size())
  97     m_vecQueue.push_back(iAdd);
  98   if (!m_bIsRunning)
  99   {
 100     StopThread();
 101     m_bIsRunning = true;
 102     CThread::Create(false, THREAD_MINSTACKSIZE);
 103   }
 104 }
 105
 106 void CRssReader::OnExit()
 107 {
 108   m_bIsRunning = false;
 109 }
 110
 111 int CRssReader::GetQueueSize()
 112 {
 113   CSingleLock lock(*this);
 114   return m_vecQueue.size();
 115 }
 116
 117 void CRssReader::Process()
 118 {
 119   while (GetQueueSize())
 120   {
 121     EnterCriticalSection(*this);
 122
 123     int iFeed = m_vecQueue.front();
 124     m_vecQueue.erase(m_vecQueue.begin());
 125
 126     m_strFeed[iFeed] = "";
 127     m_strColors[iFeed] = "";
 128
 129     CFileCurl http;
 130     http.SetUserAgent(g_settings.m_userAgent);
 131     http.SetTimeout(2);
 132     CStdString strXML;
 133     CStdString strUrl = m_vecUrls[iFeed];
 134
 135     LeaveCriticalSection(*this);
 136
 137     int nRetries = 3;
 138     CURL url(strUrl);
 139
 140     // we wait for the network to come up
 141     if ((url.GetProtocol() == "http" || url.GetProtocol() == "https") && !g_application.getNetwork().IsAvailable(true))
 142       strXML = "<rss><item><title>"+g_localizeStrings.Get(15301)+"</title></item></rss>";
 143     else
 144     {
 145       unsigned int starttime = CTimeUtils::GetTimeMS();
 146       while ( (!m_bStop) && (nRetries > 0) )
 147       {
 148         unsigned int currenttimer = CTimeUtils::GetTimeMS() - starttime;
 149         if (currenttimer > 15000)
 150         {
 151           CLog::Log(LOGERROR,"Timeout whilst retrieving %s", strUrl.c_str());
 152           http.Cancel();
 153           break;
 154         }
 155         nRetries--;
 156
 157         if (url.GetProtocol() != "http" && url.GetProtocol() != "https")
 158         {
 159           CFile file;
 160           if (file.Open(strUrl))
 161           {
 162             char *yo = new char[(int)file.GetLength()+1];
 163             file.Read(yo,file.GetLength());
 164             yo[file.GetLength()] = '\0';
 165             strXML = yo;
 166             delete[] yo;
 167             break;
 168           }
 169         }
 170         else
 171           if (http.Get(strUrl, strXML))
 172           {
 173             CLog::Log(LOGDEBUG, "Got rss feed: %s", strUrl.c_str());
 174             break;
 175           }
 176       }
 177       http.Cancel();
 178     }
 179     if ((!strXML.IsEmpty()) && m_pObserver)
 180     {
 181       // erase any <content:encoded> tags (also unsupported by tinyxml)
 182       int iStart = strXML.Find("<content:encoded>");
 183       int iEnd = 0;
 184       while (iStart > 0)
 185       {
 186         // get <content:encoded> end position
 187         iEnd = strXML.Find("</content:encoded>", iStart) + 18;
 188
 189         // erase the section
 190         strXML = strXML.erase(iStart, iEnd - iStart);
 191
 192         iStart = strXML.Find("<content:encoded>");
 193       }
 194
 195       if (Parse((LPSTR)strXML.c_str(),iFeed))
 196       {
 197         CLog::Log(LOGDEBUG, "Parsed rss feed: %s", strUrl.c_str());
 198       }
 199     }
 200   }
 201   UpdateObserver();
 202 }
 203
 204 void CRssReader::getFeed(vecText &text)
 205 {
 206   text.clear();
 207   // double the spaces at the start of the set
 208   for (int j = 0; j < m_spacesBetweenFeeds; j++)
 209     text.push_back(L' ');
 210   for (unsigned int i = 0; i < m_strFeed.size(); i++)
 211   {
 212     for (int j = 0; j < m_spacesBetweenFeeds; j++)
 213       text.push_back(L' ');
 214
 215     for (unsigned int j = 0; j < m_strFeed[i].size(); j++)
 216     {
 217       character_t letter = m_strFeed[i][j] | ((m_strColors[i][j] - 48) << 16);
 218       text.push_back(letter);
 219     }
 220   }
 221 }
 222
 223 void CRssReader::AddTag(const CStdString &aString)
 224 {
 225   m_tagSet.push_back(aString);
 226 }
 227
 228 void CRssReader::AddString(CStdStringW aString, int aColour, int iFeed)
 229 {
 230   if (m_rtlText)
 231     m_strFeed[iFeed] = aString + m_strFeed[iFeed];
 232   else
 233     m_strFeed[iFeed] += aString;
 234
 235   int nStringLength = aString.GetLength();
 236
 237   for (int i = 0;i < nStringLength;i++)
 238   {
 239     aString[i] = (CHAR) (48 + aColour);
 240   }
 241
 242   if (m_rtlText)
 243     m_strColors[iFeed] = aString + m_strColors[iFeed];
 244   else
 245     m_strColors[iFeed] += aString;
 246 }
 247
 248 void CRssReader::GetNewsItems(TiXmlElement* channelXmlNode, int iFeed)
 249 {
 250   HTML::CHTMLUtil html;
 251
 252   TiXmlElement * itemNode = channelXmlNode->FirstChildElement("item");
 253   map <CStdString, CStdStringW> mTagElements;
 254   typedef pair <CStdString, CStdStringW> StrPair;
 255   list <CStdString>::iterator i;
 256
 257   bool bEmpty=true;
 258
 259   // Add the title tag in if we didn't pass any tags in at all
 260   // Represents default behaviour before configurability
 261
 262   if (m_tagSet.empty())
 263     AddTag("title");
 264
 265   while (itemNode > 0)
 266   {
 267     bEmpty = false;
 268     TiXmlNode* childNode = itemNode->FirstChild();
 269     mTagElements.clear();
 270     while (childNode > 0)
 271     {
 272       CStdString strName = childNode->Value();
 273
 274       for (i = m_tagSet.begin(); i != m_tagSet.end(); i++)
 275       {
 276         if (!childNode->NoChildren() && i->Equals(strName))
 277         {
 278           CStdString htmlText = childNode->FirstChild()->Value();
 279
 280           // This usually happens in right-to-left languages where they want to
 281           // specify in the RSS body that the text should be RTL.
 282           // <title>
 283           //  <div dir="RTL">��� ����: ���� �� �����</div>
 284           // </title>
 285           if (htmlText.Equals("div") || htmlText.Equals("span"))
 286           {
 287             htmlText = childNode->FirstChild()->FirstChild()->Value();
 288           }
 289
 290           CStdStringW unicodeText,unicodeText2;
 291
 292           fromRSSToUTF16(htmlText, unicodeText2);
 293           html.ConvertHTMLToW(unicodeText2, unicodeText);
 294
 295           mTagElements.insert(StrPair(*i, unicodeText));
 296         }
 297       }
 298       childNode = childNode->NextSibling();
 299     }
 300
 301     int rsscolour = RSS_COLOR_HEADLINE;
 302     for (i = m_tagSet.begin();i != m_tagSet.end();i++)
 303     {
 304       map <CStdString, CStdStringW>::iterator j = mTagElements.find(*i);
 305
 306       if (j == mTagElements.end())
 307         continue;
 308
 309       CStdStringW& text = j->second;
 310       AddString(text, rsscolour, iFeed);
 311       rsscolour = RSS_COLOR_BODY;
 312       text = " - ";
 313       AddString(text, rsscolour, iFeed);
 314     }
 315     itemNode = itemNode->NextSiblingElement("item");
 316   }
 317 }
 318
 319 void CRssReader::fromRSSToUTF16(const CStdStringA& strSource, CStdStringW& strDest)
 320 {
 321   CStdString flippedStrSource;
 322   CStdString strSourceUtf8;
 323
 324   g_charsetConverter.stringCharsetToUtf8(m_encoding, strSource, strSourceUtf8);
 325   if (m_rtlText)
 326     g_charsetConverter.utf8logicalToVisualBiDi(strSourceUtf8, flippedStrSource);
 327   else
 328     flippedStrSource = strSourceUtf8;
 329   g_charsetConverter.utf8ToW(flippedStrSource, strDest, false);
 330 }
 331
 332 bool CRssReader::Parse(LPSTR szBuffer, int iFeed)
 333 {
 334   m_xml.Clear();
 335   m_xml.Parse((LPCSTR)szBuffer, 0, TIXML_ENCODING_LEGACY);
 336
 337   m_encoding = "UTF-8";
 338   if (m_xml.RootElement())
 339   {
 340     TiXmlDeclaration *tiXmlDeclaration = m_xml.RootElement()->Parent()->FirstChild()->ToDeclaration();
 341     if (tiXmlDeclaration != NULL && strlen(tiXmlDeclaration->Encoding()) > 0)
 342     {
 343       m_encoding = tiXmlDeclaration->Encoding();
 344     }
 345   }
 346
 347   CLog::Log(LOGDEBUG, "RSS feed encoding: %s", m_encoding.c_str());
 348
 349   return Parse(iFeed);
 350 }
 351
 352 bool CRssReader::Parse(int iFeed)
 353 {
 354   TiXmlElement* rootXmlNode = m_xml.RootElement();
 355
 356   if (!rootXmlNode)
 357     return false;
 358
 359   TiXmlElement* rssXmlNode = NULL;
 360
 361   CStdString strValue = rootXmlNode->Value();
 362   if (( strValue.Find("rss") >= 0 ) || ( strValue.Find("rdf") >= 0 ))
 363   {
 364     rssXmlNode = rootXmlNode;
 365   }
 366   else
 367   {
 368     // Unable to find root <rss> or <rdf> node
 369     return false;
 370   }
 371
 372   TiXmlElement* channelXmlNode = rssXmlNode->FirstChildElement("channel");
 373   if (channelXmlNode)
 374   {
 375     TiXmlElement* titleNode = channelXmlNode->FirstChildElement("title");
 376     if (titleNode && !titleNode->NoChildren())
 377     {
 378       CStdString strChannel = titleNode->FirstChild()->Value();
 379       CStdStringW strChannelUnicode;
 380       fromRSSToUTF16(strChannel, strChannelUnicode);
 381       AddString(strChannelUnicode, RSS_COLOR_CHANNEL, iFeed);
 382
 383       AddString(":", RSS_COLOR_CHANNEL, iFeed);
 384       AddString(" ", RSS_COLOR_CHANNEL, iFeed);
 385     }
 386
 387     GetNewsItems(channelXmlNode,iFeed);
 388   }
 389
 390   GetNewsItems(rssXmlNode,iFeed);
 391
 392   // avoid trailing ' - '
 393   if (m_strFeed[iFeed].size() > 3 && m_strFeed[iFeed].Mid(m_strFeed[iFeed].size()-3) == L" - ")
 394   {
 395     if (m_rtlText)
 396     {
 397       m_strFeed[iFeed].erase(0, 3);
 398       m_strColors[iFeed].erase(0, 3);
 399     }
 400     else
 401     {
 402       m_strFeed[iFeed].erase(m_strFeed[iFeed].length()-3);
 403       m_strColors[iFeed].erase(m_strColors[iFeed].length()-3);
 404     }
 405   }
 406   return true;
 407 }
 408
 409 void CRssReader::SetObserver(IRssObserver *observer)
 410 {
 411   m_pObserver = observer;
 412 }
 413
 414 void CRssReader::UpdateObserver()
 415 {
 416   if (!m_pObserver) return;
 417   vecText feed;
 418   getFeed(feed);
 419   if (feed.size() > 0)
 420   {
 421     g_graphicsContext.Lock();
 422     if (m_pObserver) // need to check again when locked to make sure observer wasnt removed
 423       m_pObserver->OnFeedUpdate(feed);
 424     g_graphicsContext.Unlock();
 425   }
 426 }
 427
 428 void CRssReader::CheckForUpdates()
 429 {
 430   SYSTEMTIME time;
 431   GetLocalTime(&time);
 432
 433   for (unsigned int i = 0;i < m_vecUpdateTimes.size(); ++i )
 434   {
 435     if (m_requestRefresh || ((time.wDay * 24 * 60) + (time.wHour * 60) + time.wMinute) - ((m_vecTimeStamps[i]->wDay * 24 * 60) + (m_vecTimeStamps[i]->wHour * 60) + m_vecTimeStamps[i]->wMinute) > m_vecUpdateTimes[i] )
 436     {
 437       CLog::Log(LOGDEBUG, "Updating RSS");
 438       GetLocalTime(m_vecTimeStamps[i]);
 439       AddToQueue(i);
 440     }
 441   }
 442
 443   m_requestRefresh = false;
 444 }
 445
 446 CRssManager g_rssManager;
 447
 448 CRssManager::CRssManager()
 449 {
 450   m_bActive = false;
 451 }
 452
 453 CRssManager::~CRssManager()
 454 {
 455   Stop();
 456 }
 457
 458 void CRssManager::Start()
 459  {
 460    m_bActive = true;
 461 }
 462
 463 void CRssManager::Stop()
 464 {
 465   m_bActive = false;
 466   for (unsigned int i = 0; i < m_readers.size(); i++)
 467   {
 468     if (m_readers[i].reader)
 469     {
 470       delete m_readers[i].reader;
 471     }
 472   }
 473   m_readers.clear();
 474 }
 475
 476 // returns true if the reader doesn't need creating, false otherwise
 477 bool CRssManager::GetReader(int controlID, int windowID, IRssObserver* observer, CRssReader *&reader)
 478 {
 479   // check to see if we've already created this reader
 480   for (unsigned int i = 0; i < m_readers.size(); i++)
 481   {
 482     if (m_readers[i].controlID == controlID && m_readers[i].windowID == windowID)
 483     {
 484       reader = m_readers[i].reader;
 485       reader->SetObserver(observer);
 486       reader->UpdateObserver();
 487       return true;
 488     }
 489   }
 490   // need to create a new one
 491   READERCONTROL readerControl;
 492   readerControl.controlID = controlID;
 493   readerControl.windowID = windowID;
 494   reader = readerControl.reader = new CRssReader;
 495   m_readers.push_back(readerControl);
 496   return false;
 497 }
 498