I am trying to write a function that pulls all url's from a string and remove a potential trailing slash from the end.
function getUrls($string) {
$regex = '/https?\:\/\/[^\" ]+/i';
preg_match_all($regex, $string, $matches);
return ($matches[0]);
}
But that returns http://test.com. (trailing period) If i have
$string = "Hi I am sharing http://test.com.";
$urls = getUrls($string);
It returns the URL with the period at the end.
http is risky.. Consider the following string: this is a comment.net is great., it would incorrectly match comment.net as a url. I recommend reading the pages linked by martineno, there is some discussion about this problem in Alan Storm's page.In case anyone comes across this, here is what I put together:
$aProtocols = array('http:\/\/', 'https:\/\/', 'ftp:\/\/', 'news:\/\/', 'nntp:\/\/', 'telnet:\/\/', 'irc:\/\/', 'mms:\/\/', 'ed2k:\/\/', 'xmpp:', 'mailto:');
$aSubdomains = array('www'=>'http://', 'ftp'=>'ftp://', 'irc'=>'irc://', 'jabber'=>'xmpp:');
$sRELinks = '/(?:(' . implode('|', $aProtocols) . ')[^\^\[\]{}|\\"\'<>`\s]*[^!@\^()\[\]{}|\\:;"\',.?<>`\s])|(?:(?:(?:(?:[^@:<>(){}`\'"\/\[\]\s]+:)?[^@:<>(){}`\'"\/\[\]\s]+@)?(' . implode('|', array_keys($aSubdomains)) . ')\.(?:[^`~!@#$%^&*()_=+\[{\]}\\|;:\'",<.>\/?\s]+\.)+[a-z]{2,6}(?:[\/#?](?:[^\^\[\]{}|\\"\'<>`\s]*[^!@\^()\[\]{}|\\:;"\',.?<>`\s])?)?)|(?:(?:[^@:<>(){}`\'"\/\[\]\s]+@)?((?:(?:(?:(?:[0-1]?[0-9]?[0-9])|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:(?:[0-1]?[0-9]?[0-9])|(?:2[0-4][0-9])|(?:25[0-5]))){3})|(?:[A-Fa-f0-9:]{16,39}))|(?:(?:[^`~!@#$%^&*()_=+\[{\]}\\|;:\'",<.>\/?\s]+\.)+[a-z]{2,6}))\/(?:[^\^\[\]{}|\\"\'<>`\s]*[^!@\^()\[\]{}|\\:;"\',.?<>`\s](?:[#?](?:[^\^\[\]{}|\\"\'<>`\s]*[^!@\^()\[\]{}|\\:;"\',.?<>`\s])?)?)?)|(?:[^@:<>(){}`\'"\/\[\]\s]+:[^@:<>(){}`\'"\/\[\]\s]+@((?:(?:(?:(?:[0-1]?[0-9]?[0-9])|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:(?:[0-1]?[0-9]?[0-9])|(?:2[0-4][0-9])|(?:25[0-5]))){3})|(?:[A-Fa-f0-9:]{16,39}))|(?:(?:[^`~!@#$%^&*()_=+\[{\]}\\|;:\'",<.>\/?\s]+\.)+[a-z]{2,6}))(?:\/(?:(?:[^\^\[\]{}|\\"\'<>`\s]*[^!@\^()\[\]{}|\\:;"\',.?<>`\s])?)?)?(?:[#?](?:[^\^\[\]{}|\\"\'<>`\s]*[^!@\^()\[\]{}|\\:;"\',.?<>`\s])?)?))|([^@:<>(){}`\'"\/\[\]\s]+@(?:(?:(?:[^`~!@#$%^&*()_=+\[{\]}\\|;:\'",<.>\/?\s]+\.)+[a-z]{2,6})|(?:(?:(?:(?:(?:[0-1]?[0-9]?[0-9])|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:(?:[0-1]?[0-9]?[0-9])|(?:2[0-4][0-9])|(?:25[0-5]))){3})|(?:[A-Fa-f0-9:]{16,39}))))(?:[^\^*\[\]{}|\\"<>\/`\s]+[^!@\^()\[\]{}|\\:;"\',.?<>`\s])?)/i';
function getUrls($string) {
global $sRELinks;
preg_match_all($sRELinks, $string, $matches);
return ($matches[0]);
}
From http://yellow5.us/journal/server_side_text_linkification/
Depending on how strict you want to be, consider the Liberal, Accurate Regex Pattern for Matching URLs regular expression pattern discussed on Daring Fireball. The pattern in full is:
\b(([\w-]+://?|www[.])[^\s()<>]+(?:\([\w\d]+\)|([^[:punct:]\s]|/)))
If you are interested in how it works, Alan Storm has a great explanation.
