[boost] Problem with subexpresions

From: John Maddock (John_Maddock_at_[hidden])
Date: 2001-04-30 06:34:02

>I expect matches[1].first to return "Moose", the first match. But it
>doesn't - it returns "Moose, Orangutang". matches[2].first returns
>"Orangutang", which makes some sense.
>
>Actually, I don't understand this whole first/second business in
>sub_matches. In any other language, matches[0] would just be a string
>or something. But here, obviously things are being done differently,
>and just don't understand what the scheme is here. I've read the docs
>over and over, and I can't find a darn thing.

The regex library is iterator based (just like the standard library
algorithms are), so when you locate a match what you get back is a pair of
*iterators* denoting the start and end of the match, and a bool flag
indicating whether that sub-expression particiapated in the match (in some
expression/text combinations some sub-expressions may be unmatched).  So
the docs define sub_match as:

template <class iterator>
struct sub_match
{
   typedef typename std::iterator_traits<iterator>::value_type      
value_type;
   typedef typename std::iterator_traits<iterator>::difference_type 
difference_type;
   typedef iterator                                                 
iterator_type;
   
   iterator first;
   iterator second;
   bool matched;

   operator std::basic_string<value_type>()const;

   bool operator==(const sub_match& that)const;
   bool operator !=(const sub_match& that)const;
   difference_type length()const;
};

It really is just a simple structure, with the range [first,second)
denoting what matched.

You can also cast the result to the string, either explicitly or
implicitly:

#include <iostream>
#include <boost\regex.hpp>

using namespace boost;
using namespace std;

int main(int argc, char* argv[])
{
        regex rxHeader("(\\w+), (\\w+)");
        cmatch matches;

        if (regex_match("Moose, Orangutang", matches, rxHeader))
        {
                // explict cast:
                cout << static_cast<std::string>(matches[1]) << endl;
                // implicit cast:
                std::string s = matches[2];
                cout << s << endl;
        }
        else
                puts("No match.");

        return 0;
}

- John Maddock
http://ourworld.compuserve.com/homepages/john_maddock/