By default preg_match_all() uses PREG_PATTERN_ORDER flag, which means:
Orders results so that $matches[0] is
an array of full pattern matches,
$matches1 is an array of strings
matched by the first parenthesized
subpattern, and so on.
See http://php.net/preg_match_all
Here is sample output:
array(
0 => array( // Full pattern matches
0 => 'http://www.w3.org/TR/html4/strict.dtd',
1 => ...
),
1 => array( // First parenthesized subpattern.
// In your case it is the same as full pattern, because first
// parenthesized subpattern includes all pattern :-)
0 => 'http://www.w3.org/TR/html4/strict.dtd',
1 => ...
),
2 => array( // Second parenthesized subpattern.
0 => 'www.w3.org',
1 => ...
),
...
)
So, as R. Hill answered, you need $matches[0] to access all matched urls.
And as budinov.com pointed, you should remove outer parentheses to avoid second match duplicate first one, e.g.:
preg_match_all('~https?://([-\w\.]+)+(:\d+)?(/([\w/_\.]*(\?\S+)?)?)?~', $content, $turls);
// where $turls[0] is what you need