Thread: PCRE string issue

I am working on PCRE for the first time . my code needs to find the @ refernce as u would see in twitter and remove it. knowing that pcre library has no replace function . i wrote myself . it works fine when i use the function once . but iterating again corrupts my string .
Any idea

Code:

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


char* sanitize(char*);
char* stringReplace(char* subj,int start,int end,char* replace);
char* pcreReplace(char* regex,char* txt);


int main()
{
    char* str="a ina @abcds http://aa.com @abcds ";
    str=sanitize(str);
    //printf("%s\n",str);
    return 0;
}


char* sanitize(char* str)
{
    //remove @ reference
    str=pcreReplace("@([^\\s]){1,}",str);
    str=pcreReplace("@([^\\s]){1,}",str);
    //str=pcreReplace("https?([^\\s]){1,}",str);
    
    return str;
}


char* pcreReplace(char* re,char* txt)
{
    const char *error;
    int erroffset;
    int ovector[186];
    
    pcre *r =  pcre_compile(re, PCRE_CASELESS|PCRE_DOTALL, &error, &erroffset, NULL);
    int rc = pcre_exec(r, NULL, txt, strlen(txt), 0, 0, ovector, 186);
    printf("%s\n",txt);
    
    if(rc<0)
        return txt;
    
    //printf("%i",ovector[0]);
    //char c1[1024];
    //pcre_copy_substring(txt, ovector, rc,1,c1, 1024);
    txt=stringReplace(txt,ovector[0],ovector[2],"");
    return txt;
}


char* stringReplace(char* subj,int start,int end,char* replace)
{
    int newLength=strlen(subj)-(end-start+1)+strlen(replace);
    char newString[newLength];
    int i;
    int itr=0;
    for(i=0;i<strlen(subj);i++)
    {
        if(i<start || i>end)
        {
            newString[itr]=subj[i];
            itr++;
        }    
    } 
    newString[newLength]='\0';
    return newString;
}

You're returning the local array (pointer) newString in stringReplace.

It's still incorrect to return a pointer to data on a function's stack since it is invalid after the function returns (and will soon be overwritten with other data).

Are you using ovector properly?

The first pair of integers, ovector[0] and ovector[1], identify the portion of the subject string matched by the entire pattern. The next pair is used for the first capturing subpattern, and so on.

Just use a loop. Here's an example.

Code:

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

void print(char *str, int start, int end) {
    while (start <= end)
        putchar(str[start++]);
    putchar('\n');
}

int main(void) {
    char str[] = "@abcd whatever @ef @ghi stuff";
    int errOffset;
    int rc;
    const char *errText;
    int ovector[6];
    pcre *re;
    int offset = 0;

    re = pcre_compile("@\\S+", 0, &errText, &errOffset, NULL);
    if (re == NULL) {
        fprintf(stderr, "Error: %s (%d)\n", errText, errOffset);
        exit(1);
    } else {
        while (1) {
            rc = pcre_exec(re, NULL, str, strlen(str), offset, 0, ovector, 6);
            if (rc <= 0)
                break;
            print(str, ovector[0], ovector[1]);
            offset = ovector[1] + 1;  // move offset past current match
        }
        pcre_free(re);
    }

    return 0;
}