Regex Match PHP Comment

Question

Ive been trying to match PHP comments using regex.

//([^<]+)\r\n

Thats what ive got but it doesn't really work.

Ive also tried

//([^<]+)\r
//([^<]+)\n
//([^<]+)

...to no avail

Answer 1

You probably need to escape the "//":

\/\/([^<]+)

Answer 2

In what program are you coding this regex? Your final example is a good sanity check if you're worried that the newline chars aren't working. (I have no idea why you don't allow less-than in your comment; I'm assuming that's specific to your application.)

Try

//[^<]+

and see if that works. As Draemon says, you might have to escape the diagonals. You might also have to escape the parentheses. I can't tell if you know this, but parentheses are often used to enclose capturing groups. Finally, check whether there is indeed at least one character after the double slashes.

Answer 3

To match comments, you have to think there are two types of comments in PHP 5 :

• comments which start by // and go to the end of the line
• comments that start by /* and go to */

Considering you have these two lines first :

$filePath = '/home/squale/developpement/astralblog/website/library/HTMLPurifier.php';
$str = file_get_contents($filePath);

You could match the first ones with :

$matches_slashslash = array();
if (preg_match_all('#//(.*)$#m', $str, $matches_slashslash)) {
    var_dump($matches_slashslash[1]);
}

And the second ones with :

$matches_slashstar = array();
if (preg_match_all('#/\*(.*?)\*/#sm', $str, $matches_slashstar)) {
    var_dump($matches_slashstar[1]);
}

But you will probably get into troubles with '//' in the middle of string (what about heredoc syntax, btw, did you think about that one ? ), or "toggle comments" like this :

/*
echo 'a';
/*/
echo 'b';
//*/

(Just add a slash at the begining to "toggle" the two blocks, if you don't know the trick)

So... Quite hard to detect comments with only regex...

---

Another way would be to use the PHP Tokenizer, which, obviously, knows how to parse PHP code and comments.

For references, see :

• token_get_all
• List of Parser Tokens

With that, you would have to use the tokenizer on your string of PHP code, iterate on all the tokens you get as a result, and detect which ones are comments.

Something like this would probably do :

$tokens = token_get_all($str);

foreach ($tokens as $token) {
    if ($token[0] == T_COMMENT
        || $token[0] == T_DOC_COMMENT) {
        // This is a comment ;-)
        var_dump($token);
    }
}

And, as output, you'll get a list of stuff like this :

array
  0 => int 366
  1 => string '/** Version of HTML Purifier */' (length=31)
  2 => int 57

or this :

array
  0 => int 365
  1 => string '// :TODO: make the config merge in, instead of replace
' (length=55)
  2 => int 117

(You "just" might to strip the // and /* */, but that's up to you ; at least, you have extracted the comments ^^ )

If you really want to detect comments without any kind of strange error due to "strange" syntax, I suppose this would be the way to go ;-)