How to limit my regex that is detecting too much? - regex

Question

I have a regex that is attempting to detect title & link markup:
[title](http://link.com)
So far I have:
(\[)(.*?)(\])(\(((http[s]?)|ftp):\/\/)(.*?)(\))
Which is detecting to much when an untitled link markup is before it
[http://google.com] [Digg](http://digg.com)
[Internal Page] Random other text [Digg](http://digg.com)
How can I limit the regex to just the titled link?
Full PHP for titled & untitled links:
// Titled Links
// [Digg](http://digg.com)
// [Google](http://google.com)
$text = preg_replace_callback(
'/(\[)(.*?)(\])(\(((http[s]?)|ftp):\/\/)(.*?)(\))/',
function ($match) {
$link = trim($match[7]);
$ret = "<a target='_blank' href='" . strtolower($match[5]) . "://" . $link . "'>" . trim($match[2]) . "</a>";
if (strtolower($match[5]) == "http") {
$ret .= "<img src='/images/link_http.png' class='link' />";
} else if (strtolower($match[5]) == "https") {
$ret .= "<img src='/images/link_https.png' class='link' />";
} else if (strtolower($match[5]) == "ftp") {
$ret .= "<img src='/images/link_ftp.png' class='link' />";
}
return $ret;
},
$text
);
// Untitled Links
// [Internal Page]
// [http://google.com]
$text = preg_replace_callback(
'/(\[)(.*?)(\])/',
function ($match) {
$link = trim($match[2]);
$ret = "";
if ($this->startsWith(strtolower($link), "https")) {
$ret = "<a target='_blank' href='" . $link . "'>" . $link . "</a>";
$ret .= "<img src='/images/link_https.png' class='link' />";
} else if ($this->startsWith(strtolower($link), "http")) {
$ret = "<a target='_blank' href='" . $link . "'>" . $link . "</a>";
$ret .= "<img src='/images/link_http.png' class='link' />";
} else if ($this->startsWith(strtolower($link), "ftp")) {
$ret = "<a target='_blank' href='" . $link . "'>" . $link . "</a>";
$ret .= "<img src='/images/link_ftp.png' class='link' />";
} else {
$link = str_replace(" ", "_", $link);
$ret = "<a href='" . $link . "'>" . trim($match[2]) . "</a>";
}
return $ret;
},
$text
);

Answer 1

If you're trying to go through Markdown links, you'll probably want to grab the regex and logic straight from the source:
https://github.com/michelf/php-markdown/blob/lib/Michelf/Markdown.php#L510
https://github.com/tanakahisateru/js-markdown-extra/blob/master/js-markdown-extra.js#L630

Answer 2

Make the title optional by appending a '?' to the group that matches the title.

Answer 3

Instead of (.*?) try matching on something you really don't want, like a space e.g. ([^\s]+).
Also, the whole of the second part is optional (if you can have a untitled link), so add the ? as #Arnout suggests, e.g.
(\(((http[s]?)|ftp):\/\/)([^\s]+)(\))?
May I also suggest, (though I'm not sure it's supported in PHP regex it appears to be, using the whitespace flag and breaking it up over a few lines for readability:
/
(
\[
)
(.*?)
(
\]
)
(
\(
(
(http[s]?)
|
ftp
)
:\/\/
)
(.*?)
(
\)
)
/x
That is a lot clearer, and it's easier to see:
The [s]? could just be s?
The scheme brackets should be ((?:https?)|(?:ftp)) or it only ORs on the f, and you get additional unneeded captures.
You may possibly be able to comment on it too, within the regex (again, I'm unsure you can with PHP).

Answer 4

This works but doesn't have the parentheses for all the groups you're trying to match.
\[[\w\s]+\]\((https?|ftp)://[^)]+\)