URL regex with regex.h in c

Question

SO:

I'm having a bit of difficulty setting up a regex to match a URL using the regex.h library in c. I have a working IP regex that I was hoping to convert to match a simple string such as www.alphanumerictext12.com|edu|org. Something is wrong with my syntax in the regex definition itself.

Below is the working IPREGEX code and my attempted URL regex.

#define IPEXPR    "([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})"
#define URLEXPR "(www)\\.((?:[a-z][a-z]*[0-9]+[a-z0-9]*))\\.(com|edu|org)"

regex_t regex;

if(regcomp(&regex, IPEXPR, REG_EXTENDED) != 0)
  return 0;

if(regexec(&regex, inputURL, 0, NULL, 0) != 0)
  return 0;

Thanks all for your help. cb

Answer 1

Try:

"www\\.[a-z]+[a-z0-9]*\\.(com|edu|org)"

I removed the [0-9]+ and replaced [a-z][a-z]* with [a-z]+.

Answer 2

Your slashes are going the wrong way. Remember the web uses forward slashes.

Answer 3

The problem is in (?: ), You need just (www)\\.([a-z][a-z]*[0-9]+[a-z0-9]*)\\.(com|edu|org).

Btw, your inner expression says: "at least one alpha character, then at least one numeric character, then any alphanumeric characters". Is it what you mean? If so, you can make it a little bit shorter: [a-z]+[0-9]+[a-z0-9]*.

Answer 4

From Coding Horror:

Some people, when confronted with a problem, think "I know, I'll use regular expressions." Now they have two problems.

What I mean is: are you sure a regular expression is the best way to solve your problem? Maybe you can test whether the string is a URL with some more lightweigth method?

---

Edit

The following program on my computer, with output redirected to /dev/null, prints (to stderr)

rx time: 1.730000
lw time: 0.920000

Program Listing:

#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <regex.h>
#include <string.h>
#include <time.h>

int goodurl_rx(const char *buf) {
  static regex_t rx;
  static int done = 0;
  int e;

  if (!done) {
    done = 1;
    if ((e = regcomp(&rx, "^www\\.[a-z][a-z0-9]*\\.(com|edu|org)$", REG_EXTENDED)) != 0) {
      printf("Error %d compiling regular expression.\n", e);
      exit(EXIT_FAILURE);
    }
  }
  return !regexec(&rx, buf, 0, NULL, 0);
}

int goodurl_lw(const char *buf) {
  if (*buf++ != 'w') return 0;
  if (*buf++ != 'w') return 0;
  if (*buf++ != 'w') return 0;
  if (*buf++ != '.') return 0;
  if (!isalpha((unsigned char)*buf++)) return 0;
  while (isalnum((unsigned char)*buf)) buf++;
  if (*buf++ != '.') return 0;
  if ((*buf == 'c') && (*(buf+1) == 'o') && (*(buf+2) == 'm') && (*(buf+3) == 0)) return 1;
  if ((*buf == 'e') && (*(buf+1) == 'd') && (*(buf+2) == 'u') && (*(buf+3) == 0)) return 1;
  if ((*buf == 'o') && (*(buf+1) == 'r') && (*(buf+2) == 'g') && (*(buf+3) == 0)) return 1;
  return 0;
}

int main(void) {
  clock_t t0, t1, t2;
  char *buf[] = {"www.alphanumerics.com", "ww2.alphanumerics.com", "www.alphanumerics.net"};
  int times;

  t0 = clock();
  times = 1000000;
  while (times--) {
    printf("    %s: %s\n", buf[0], goodurl_rx(buf[0])?"pass":"invalid");
    printf("    %s: %s\n", buf[1], goodurl_rx(buf[1])?"pass":"invalid");
    printf("    %s: %s\n", buf[2], goodurl_rx(buf[2])?"pass":"invalid");
  };
  t1 = clock();
  times = 1000000;
  while (times--) {
    printf("    %s: %s\n", buf[0], goodurl_lw(buf[0])?"pass":"invalid");
    printf("    %s: %s\n", buf[1], goodurl_lw(buf[1])?"pass":"invalid");
    printf("    %s: %s\n", buf[2], goodurl_lw(buf[2])?"pass":"invalid");
  } while (0);
  t2 = clock();

  fprintf(stderr, "rx time: %f\n", (double)(t1-t0)/CLOCKS_PER_SEC);
  fprintf(stderr, "lw time: %f\n", (double)(t2-t1)/CLOCKS_PER_SEC);
  return 0;
}

Answer 5

You probably should be using inet_pton() which is a standard POSIX function (replacing inet_aton()) and handles both IPv4 and IPv6 address formats.