C++ regex string capture

Question

Tring to get C++ regex string capture to work. I have tried all four combinations of Windows vs. Linux, Boost vs. native C++ 0x11. The sample code is:

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

using namespace std;
using namespace boost;

int main(int argc, char** argv)
{
    smatch sm1;
    regex_search(string("abhelloworld.jpg"), sm1, regex("(.*)jpg"));
    cout << sm1[1] << endl;
    smatch sm2;
    regex_search(string("hell.g"), sm2, regex("(.*)g"));
    cout << sm2[1] << endl;
}

The closest that works is g++ (4.7) with Boost (1.51.0). There, the first cout outputs the expected abhelloworld. but nothing from the second cout.

g++ 4.7 with -std=gnu++11 and <regex> instead of <boost/regex.hpp> produces no output.

Visual Studio 2012 using native <regex> yields an exception regarding incompatible string iterators.

Visual Studio 2008 with Boost 1.51.0 and <boost/regex.hpp> yields an exception regarding "Standard C++ Libraries Invalid argument".

Are these bugs in C++ regex, or am I doing something wrong?

bobobobo · Accepted Answer · 2021-09-08 22:54:59Z

Are these bugs in C++ regex, or am I doing something wrong?

At the time of your posting, gcc didn't support <regex> as noted in the other answer (it does now). As for the other problems, your problem is you are passing temporary string objects. Change your code to the following:

smatch sm1;
string s1("abhelloworld.jpg");
regex_search(s1, sm1, regex("(.*)jpg"));
cout << sm1[1] << endl;
smatch sm2;
string s2("hell.g");
regex_search(s2, sm2, regex("(.*)g"));
cout << sm2[1] << endl;

Your original example compiles because regex_search takes a const reference which temporary objects can bind to, however, smatch only stores iterators into your temporary object which no longer exists. The solution is to not pass temporaries.

If you look in the C++ standard at [§ 28.11.3/5], you will find the following:

Returns: The result of regex_search(s.begin(), s.end(), m, e, flags).

What this means is that internally, only iterators to your passed in string are used, so if you pass in a temporary, iterators to that temporary object will be used which are invalid and the actual temporary itself is not stored.

Rapptz · Accepted Answer · 2012-08-25 19:39:50Z

0

GCC doesn't support <regex> yet. Refer to the Manual

answered Aug 25, 2012 at 19:39

Rapptz

21.4k5 gold badges74 silver badges87 bronze badges

1 Comment

Nicolas Holthaus Over a year ago

as of gcc4.9, c++11 <regex> is supported.

Collectives™ on Stack Overflow

C++ regex string capture

2 Answers 2

Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related