Problem with regexp in c

alitrix · 09-06-2007, 09:08 AM

I have the next code:

Code:

#include <stdio.h>
#include <regex.h>
 
int main () {
        int MAX_MATCH=5;
        regmatch_t pmatch[MAX_MATCH];
        int regflags=REG_ICASE|REG_EXTENDED;
        regex_t regy;
 
        char *ptr;
        char *regnow = "\\wtest\\w";
        regcomp(&regy, regnow, regflags);
 
        char *cleanstr = "this is a test oke?";
 
        if (regexec(&regy, cleanstr, MAX_MATCH, pmatch, 0)==REG_NOMATCH) {
                printf("No match found for: %s\n", cleanstr);
        } else {
                printf("Match found! %s\n",cleanstr);
                printf("start: %d - end: %d\n", pmatch[0].rm_so, pmatch[0].rm_eo);
        }
 
        return 0;
}

But I'm getting No match found.
It's like every \ is ignored.
If I try \\btest\\b, it doesn't work as well.

What am I doing wrong

?

Greetz,

alitrix

theNbomr · 09-06-2007, 09:17 AM

You should test the return value from regcomp(). It may be complaining about:

Code:

char *regnow = "\\wtest\w";
                       ^

... which seems to be missing one backslash.

--- rod.

alitrix · 09-06-2007, 09:29 AM

It isn't that

That was just a typo when I copyed the code to here and modified it.
Even with \\wtest\\w it's not working :s

theNbomr · 09-06-2007, 09:35 AM

Still, you should test the return value from all applicable functions. They return a status value to help you deal with errors. If you ignore what it is telling you, you will not be able to handle errors. There is even the regerror() function that makes the error messages more human readable.
--- rod.

alitrix · 09-06-2007, 09:56 AM

Code:

#include <stdio.h>
#include <regex.h>

int main () {
        int MAX_MATCH=5;
        regmatch_t pmatch[MAX_MATCH];
        int regflags=REG_ICASE|REG_EXTENDED;
        regex_t regy;

        char *ptr;
        char *regnow = "\\wtest\\w";
        regcomp(&regy, regnow, regflags);

        char *cleanstr = "this is a test oke?";
        int error;
        int maxLen=512;
        char *errorBuf = malloc(maxLen);
        memset(errorBuf, 0, maxLen);

        if ((error = regexec(&regy, cleanstr, MAX_MATCH, pmatch, 0))==REG_NOMATCH) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("No match found for %d: %s\n", error, cleanstr);
                printf("Error?: %s\n", errorBuf);
        } else {
                printf("Match found! %s\n",cleanstr);
                printf("start: %d - end: %d\n", pmatch[0].rm_so, pmatch[0].rm_eo);
        }

        return 0;
}

Output:

Code:

No match found for 1: this is a test oke?
Error?: regexec() failed to match

theNbomr · 09-06-2007, 11:17 AM

How do you know whether the regex was properly compiled? You still haven't tested the return from regcomp().
--- rod.

alitrix · 09-07-2007, 06:10 AM

Well, the regex is compiled correctly.

Code:

#include <stdio.h>
#include <regex.h>

int main () {
        int MAX_MATCH=5;
        regmatch_t pmatch[MAX_MATCH];
        int regflags=REG_ICASE|REG_EXTENDED;
        regex_t regy;

        char *ptr;
        char *regnow = "\\wtest\\w";
        int maxLen=512;
        char *errorBuf = malloc(maxLen);
        int error;

        error = regcomp(&regy, regnow, regflags);
        printf("Compile regexp? %d\n",error);
        if (error) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("Error1?: %s\n", errorBuf);
        }

        char *cleanstr = "this is a test oke?";
        memset(errorBuf, 0, maxLen);

        if ((error = regexec(&regy, cleanstr, MAX_MATCH, pmatch, 0))==REG_NOMATCH) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("No match found for %d: %s\n", error, cleanstr);
                printf("Error2?: %s\n", errorBuf);
        } else {
                printf("Match found! %s\n",cleanstr);
                printf("start: %d - end: %d\n", pmatch[0].rm_so, pmatch[0].rm_eo);
        }

        return 0;
}

Output

Code:

Compile regexp? 0
No match found for 1: this is a test oke?
Error2?: regexec() failed to match

sitycent · 09-08-2007, 06:30 AM

Quote:

Originally Posted by alitrix

Well, the regex is compiled correctly.

Code:

#include <stdio.h>
#include <regex.h>

int main () {
        int MAX_MATCH=5;
        regmatch_t pmatch[MAX_MATCH];
        int regflags=REG_ICASE|REG_EXTENDED;
        regex_t regy;

        char *ptr;
        char *regnow = "\\wtest\\w";
        int maxLen=512;
        char *errorBuf = malloc(maxLen);
        int error;

        error = regcomp(&regy, regnow, regflags);
        printf("Compile regexp? %d\n",error);
        if (error) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("Error1?: %s\n", errorBuf);
        }

        char *cleanstr = "this is a test oke?";
        memset(errorBuf, 0, maxLen);

        if ((error = regexec(&regy, cleanstr, MAX_MATCH, pmatch, 0))==REG_NOMATCH) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("No match found for %d: %s\n", error, cleanstr);
                printf("Error2?: %s\n", errorBuf);
        } else {
                printf("Match found! %s\n",cleanstr);
                printf("start: %d - end: %d\n", pmatch[0].rm_so, pmatch[0].rm_eo);
        }

        return 0;
}

Output

Code:

Compile regexp? 0
No match found for 1: this is a test oke?
Error2?: regexec() failed to match

char *cleanstr = "this is a test oke?";
memset(errorBuf, 0, maxLen

theNbomr · 09-08-2007, 10:56 AM

Okay, after building and running your code, a few observations.

For malloc() you should include malloc.h
For memset(), you should include string.h

The biggy, and I can't believe I didn't pick up on this, is the regex. I think the metachars should be \s, not \w, since you are trying to match whitespace, not word chars. This worked for me.

Code:

char *regnow = "\\stest\\s";

--- rod.

alitrix · 09-12-2007, 02:31 AM

Strange, can u show me your code?
It isn't working here

This is what I have:

Code:

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

int main () {
   int MAX_MATCH=5;
   regmatch_t pmatch[MAX_MATCH];
   int regflags=REG_ICASE|REG_EXTENDED;
   regex_t regy;

   char *ptr;
   char *regnow = "\\stest\\s";
   int maxLen=512;
   char *errorBuf = malloc(maxLen);
   int error;

   error = regcomp(&regy, regnow, regflags);
   printf("Compile regexp? %d\n",error);
   if (error) {
      regerror(error, &regy, errorBuf, maxLen);
      printf("Error1?: %s\n", errorBuf);
   }   

   char *cleanstr = "this is a test oke?";
   memset(errorBuf, 0, maxLen);

   if ((error = regexec(&regy, cleanstr, MAX_MATCH, pmatch, 0))==REG_NOMATCH) {
      regerror(error, &regy, errorBuf, maxLen);
      printf("No match found for %d: %s\n", error, cleanstr);
      printf("Error2?: %s\n", errorBuf);
   } else {
      printf("Match found! %s\n",cleanstr);
      printf("start: %d - end: %d\n", pmatch[0].rm_so, pmatch[0].rm_eo);
   }   

   return 0;
}

Output

Code:

./regex 
Compile regexp? 0
No match found for 1: this is a test oke?
Error2?: regexec() failed to match

theNbomr · 09-12-2007, 08:12 AM

Sure. It's really unchanged, except those changes that I mentioned.

Code:

[bomr@localhost c]$ more regexTest.c
#include <stdio.h>
#include <malloc.h>
#include <string.h>
#include <regex.h>

int main () {
        int MAX_MATCH=5;
        regmatch_t pmatch[MAX_MATCH];
        int regflags=REG_ICASE|REG_EXTENDED;
        regex_t regy;

        char *ptr;
        char *regnow = "\\stest\\s";
        int maxLen=512;
        char *errorBuf = malloc(maxLen);
        int error;

        error = regcomp(&regy, regnow, regflags);
        printf("Compile regexp? %d\n",error);
        if (error) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("Error1?: %s\n", errorBuf);
        }

        char *cleanstr = "this is a test oke?";
        memset(errorBuf, 0, maxLen);

        if ((error = regexec(&regy, cleanstr, MAX_MATCH, pmatch, 0))==REG_NOMATCH) {
                regerror(error, &regy, errorBuf, maxLen);
                printf("No match found for %d: %s\n", error, cleanstr);
                printf("Error2?: %s\n", errorBuf);
        } else {
                printf("Match found! %s\n",cleanstr);
                printf("start: %d - end: %d\n", pmatch[0].rm_so, pmatch[0].rm_eo);
        }

        return 0;
}

The runtime results:

Code:

[bomr@localhost c]$ ./regexTest
Compile regexp? 0
Match found! this is a test oke?
start: 9 - end: 15

Built on Fedora Core 5: gcc (GCC) 4.1.1 20070105 (Red Hat 4.1.1-51)

--- rod.

alitrix · 09-12-2007, 08:28 AM

How do u compile it?

I do: gcc regexTest.c -o regexTest

theNbomr · 09-12-2007, 10:15 AM

I use the default make rules:

Code:

make regexTest

It echoes:

Code:

cc     regexTest.c   -o regexTest

cc is a symbolic link to gcc. Looks pretty similar to your command.
--- rod.