[Boost-users] using boost.regex on submatches.

From: Dan Smithers (dsmithers_at_[hidden])
Date: 2008-05-14 04:27:45

Hi,
I have a problem using boost.regex to parse a message.
The message contains an unknown number of handles terminated by a tab.
Each handle has four hex digits. To find the block of handles I use

a regex search and then I want to use an iterator to split up the block.

#include <iostream>
#include <boost/regex.hpp>
#include <string>

int main (int argc, char *argv[])
{
  std::string input("HEAD123456789abcdef0\t");

  try {
    std::cout << "creating regex" << std::endl;
    boost::regex reg("^HEAD(([a-fA-F0-9]{4})+)\\t");
    boost::smatch sm;
    std::cout << "searching " << input << std::endl;
    boost::regex_search(input, sm, reg);

    boost::sregex_iterator end;
    boost::regex handles("([a-fA-F0-9]{4})");

    boost::sregex_iterator it(sm[1].str().begin(), sm[1].str().end(),
handles);
    while (it != end)
      {
        std::cout << *it++ << std::endl;
      }
  }
  catch (...)
    {
      std::cout << "Exception" << std::endl;
    }
}

I get a segmentation fault if I run this - sometimes with a memory
exhausted notice.

If I replace the iterator line

boost::sregex_iterator it(sm[1].str().begin(), sm[1].str().end(), handles);

with

const string& hs(sm[1].str());
boost::sregex_iterator it(hs.begin(), hs.end(), handles);

then everything seems to work.

What's going on here? Is there a simpler way of achieving this?

thanks

dan