gistfile1.cpp

#include <fstream>
#include <iostream>
#include <cstdio>
#include <cstring>

#include <htmlcxx/html/ParserDom.h>
#include <htmlcxx/html/utils.h>

int main(int argc, char** argv) 
{
  std::ifstream file(argv[1]);
  std::string html;
  while (file.good())
  {
    char buf[1024];
    memset(buf, 0, 1024);
    file.read(buf, 1024);
    html.append(buf, file.gcount());
  }
  file.close();

  htmlcxx::HTML::ParserDom parser;
  parser.parse(html);

  tree<htmlcxx::HTML::Node> root = parser.getTree();

  bool print = false;
  int depth = 0;
  tree<htmlcxx::HTML::Node>::iterator it = root.begin();
  for (;it!=root.end();++it)
  {
    if (it->tagName() == "div")
    {
      it->parseAttributes();

      if (it->attribute("class").second.find("entry ") != std::string::npos)
      {
        depth = root.depth(it);
        print = true;
      }
      else if (depth == root.depth(it))
      {
        depth = 0;
        print = false;
      }
    }

    if (print)
    {
      std::cout << it->tagName() << std::endl;
    }
  }

  return(0);
}