need reg expr help plz

zmola

hi. i'm trying to make a reg expr that will match e-mail addresses at a web site. these addresses are between <div class="lit"> and </div> tags, and they also are in the form of either (i) domain.ext or (ii) @domain.ext

using just the div tags picks up all text between the tags. i just want the domain.ext part and have been unsuccessful. help?

<?
$file = fopen ("http://www.website.com", "r");
while (!feof ($file)) {
    $line = fgets ($file, 1024);
//    if (eregi ("<div class=\"lit\">(.*)</div>", $line, $out)) {
//    if (eregi ("([a-z0-9]+([\.-][a-z0-9]+)*)+\\.[a-z]{2,4}", $line, $out)) {
//    help me please
        $email[] .= $out[1];
    }
}
fclose($file);
for ($i=0; $i < count($email); $i++) {
echo $email[$i] . "<br>";
}
?>

bbaassiri

what version of php are you using

Note: Up to (and including) PHP 4.1.0 $regs will be filled with exactly ten elements, even though more or fewer than ten parenthesized substrings may actually have matched. This has no effect on ereg()'s ability to match more substrings. If no matches are found, $regs will not be altered by ereg().

archangel_617b

http://regexlib.com/

All your regex needs. Just search on "email" or "email validation" and a couple of gems will show right up. I currently use the first one listed and it works wonderfully.

zmola

thanks for the comments and link. i did figure out how to do this. i incorporated a do-while loop to fetch anything of the form ([a-z0-9]+([.-][a-z0-9]+)*)+\.[a-z]{2,6} and stuck in a flag so i get all such occurrences that are on a line. then i just print out such occurrences as long as they are not prefaced with [url]http://[/url] and do not start with www and do end in a valid domain extension.

ednark

quote+\.[a-z]{2,6}[/quote]

ok this is my understanding: if i have anything wrong i
d be the first to want to here so don't be shy

as well as just letters and numbers these are a list of acceptable characters in the user part (and domain for that matter) as identified in rfc 2822 minus any email comments... each domain gets to restrict the acceptable characters for their email usernames though, so you won't really find &&&@^{^{^.org}} though it's technically a legal email...

a-zA-Z0-9!#\$\%\&\'*+-\/\=\?^_`{|}\~

and there is no sense restricting the domain to {2,6}

as there are allowable 1 character top level domains (though they are reserved by icann but could become active sometime) and while the 6 takes care of museum they might just pop another one iwth 7 out someday... then someone will wonder why your script doesn't work

Hikari

Like ednark said, the problem with validation scripts are new top level domains could pop up out of no where. So check out the getmxrr function. It gets the mx record for the specific domain, seeing if there is indeed an email server running at the particular destination. Try this somewhere in your script

      $domain = substr(strstr($email[$i],"@"),1)) {
        $temp = getmxrr($domain, $mx);
        if ($mx[0]) {
          echo $email[$i];
         }      

      }

That should work if you just throw it into your script

zmola

thanks for the comments and snippet. actually, my script is not a validation script. i am using it to compile a list of free email addresses. these addresses are embedded in html and are of the form domain.ext, @domain.ext, subdomain.domain.ext, etcetera, with no user names and @ isn't necessarily present. occasionally i'll get something fake like document.mm but for the most part it seems to work okay, but there's probably a better way to do this.