[Subway]

I'm working on a php application that scans through php files (phpPackageScan). Here's a sample output of a scan (WordPress 1.5): http://www.osxcode.com/phpPackageScans/wordpress/

One of the things the script does before it searches for functions and cross references is strip comments. This is the regex I'm using:

PHP Code:

$content = preg_replace('@(/\*.*?\*/)@se', "remove_non_linebreaks('\\1',$step)", $content);

                        

$content = preg_replace('/(#+[^\n]*?\n)/se', "remove_non_linebreaks('\\1',$step)", $content);

                        

$content = preg_replace('@(//+([^\n]*?\n))@se', "remove_non_linebreaks('\\1',$step)", $content); 

That function removes the comments without removing the linebreaks. Now the problem is that if the php file contains something like:

PHP Code:

if (preg_match('#^/*' . $this->index . '#', $this->permalink_structure)) { 

And another /* */ comment somewhere below, it will strip everything between even if it isn't a comment ... obviously not a good thing.

In the above linked scan, you can see the problem if you view the comment of the rewrite_rules function inside /wp-includes/classes.php

Has anyone an idea how I could make that regex act more perfect? If possible without making the script too slow?

Fredi