Thread: Problem with regex expression

Hi all,

I have a problem with a regular expression in c++. It should be very simple to do, but I don't find any helpful information on the web regarding the usage of the regex.h library. I don't understand why my pattern should be wrong. Maybe someone can explain me where I'm wrong.

I have to use the regex.h library becausing I am using it on an embedded system where I don't want to use some other library like boost.

Code:

#include "stdio.h"
#include "stdlib.h"
#include "regex.h"
#include <string>

using std::string;

int main() {
    string str = "ab:";
    // this doesn't work
    string pattern = "[0-9|a-f][0-9|a-f]:";
    // this doesn't work either
    //string pattern = "[0-9|a-f]{2}:";
    // this would work - but it's not the correct pattern
    //string pattern = "[0-9|a-f]*[0-9|a-f]*:";
    regex_t pattern_t;

    // I have also used this
    //     int ret = regcomp(&pattern_t, pattern.c_str(), 0);
    int ret = regcomp(&pattern_t, pattern.c_str(), REG_EXTENDED|REG_NOSUB);
    printf ("ret: %d\n", ret);
    ret = regexec(&pattern_t, pattern.c_str(), (size_t) 0, NULL, 0);
    printf ("ret: %d\n", ret);

    return 0;

    // correct output
    // ret: 0
    // ret: 0
}

I would be greateful for any help or hint.

pippo

Ok, I did have definitely a very simple programming error. This now is the correct version of the code. Maybe it's helpful for someone else which gets stuck into regular expressions using regex.h

Code:

#include "stdio.h"
#include "stdlib.h"
#include "regex.h"
#include <string>

using std::string;

int main() {
    string str = "Ae:3B:22:77:AC:22";
    string pattern = "[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}";
    regex_t pattern_t;
    regmatch_t pmatch_t;

    int ret = regcomp(&pattern_t, pattern.c_str(), REG_ICASE|REG_EXTENDED);
    printf ("ret: %d\n", ret);
    ret = regexec(&pattern_t, str.c_str(), 1, &pmatch_t, 0);
    printf("ret: %d\n", ret);

    if (ret == 0 && pmatch_t.rm_so == 0 && str.length() == 17)
    {
        printf("Correct MAC address!\n");
    }
    else
    {
        printf("Wrong MAC address!\n");
    }

    regfree(&pattern_t);

    return 0;
}

I want to check if a given string represents a MAC address. That's it. Thx anyway for your help.

As far as I got known and tested regex.h doesn't support nested regular expressions. First I wanted to write

Code:

string pattern = "[[[:xdigit:]]{2}:]{5}[[:xdigit:]]{2}";

but that doesn't provide the correct results. That's why I have to use the much longer version

Code:

string pattern = "[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}:[[:xdigit:]]{2}";

which is definitely not so smart but at least works

pippo

You say that like it is as a good idea to use a '_t' suffix on type names.

Soma

The version with parenthesis works like a shame. Perfect!

Code:

([[:xdigit:]]{2}:){5}[[:xdigit:]]{2}

In relation to the coding standard it seems obviously better not to use the '_t' suffix on variable names. I will respect that in the future.

Thx
pippo