5 replies to this topic

[rocketboy9000]

Regexes are an extremely useful feature found in many higher level interpreted languages, like Python, Perl, Ruby, and PHP. But with the standard posix header "regex.h", you can have them in regular ol' C! How lovely. I'll assume you know how to write them, so let's get down to business. Say we want to match big words (at least ten letters) that ends in "tion". We write the following regex into a string. We're using extended regular expressions which have the more fancy syntax we've come to know and love.

const char *pattern="[a-zA-Z]{6,}tion";

Now, how to use this? Well, it turns out that first we have to compile it into a structure called a regex_t, using the function regcomp.

regex_t rx;
regcomp(&rx,pat,REG_EXTENDED);

We must also make a variable of type regmatch_t to hold the data we'll get from the regex. We then call regexec to match the pattern to a string:

char s[100];
fgets(s,100,stdin);
regmatch_t res;
int matched regexec(&rx,s,1,&res,0);

Now, matched will be 0 if it matched, or E_NOMATCH if it didn't. res.rm_so contains the coordinate of the start of the matching substring if any, and res.rm_eo contains the end coordinate.
You can call regexec over and over on different strings. Eventually, though, you'll want to stop using the regex. Then you should call regfree to free the memory used by the compiled regex.

regfree(&rx);

To demonstrate, here is a function which matches a string to a pattern and returns a copy of the result.

regmatch_t rxmatch(char *s,char *pat){//returns a dst copy of matching substring.
    regex_t rx;
    regmatch_t res;
    regcomp(&rx,pat,REG_EXTENDED);
    regexec(&rx,s,1,&res,0);
    regfree(&rx);
    return res;
}

Generally though you should compile each regex you use only once, especially if you use a single regex on many strings.

Edited by Alexander, 03 December 2011 - 01:56 AM.

[Alexander]

Ah, a taste of regular expressions for POSIX - a great starter. Thank you.

An ironic thing, I had used some regular expressions to translate [tt] in to [noparse][inline][/noparse] tags in your post that we have available instead!

Now that the tutorial is "regexes in C", do you know what other operating system's (i.e. Windows) headers for regular expressions are?

[TheCompBoy]

Great tutorial! Very usefull.

[DarkLordofthePenguins]

I thought regex.h was deprecated.

[rocketboy9000]

Nope that's regexp.h

[fread]

Vary handy, gave me a couple of ideas.