Extracting machine tags from a string.

gistfile1.php

<?php

// Namespace and predicate start with a letter or underscore,
// followed by any number of letters, numbers, or underscores.
// If value contains spaces, it needs to be wrapped in quotes.
// Literal quotes within quotes need to be escaped using a backslash (\)


$test_cases='
#repost
#repost:url=http://example.com/post/2
#repost:content=\'Hello World.
Did I hear you say "Hello"
How are you doing \\\'Foobar\\\'?\'
#repost:published="3 hrs \'ago\'"
#repost:author="Foo \\"Baz\\" Bar"
#repost:author_url=http://example.com/foobar/
#repost:author_photo=http://example.com/foobar/photo

Non-machine-tag content
  #machinetags
';

function extract_machine_tags($str)
{
	preg_match_all('/#(?P<namespace>[a-zA-Z_][a-zA-Z0-9_]+):(?P<predicate>[a-zA-Z_][a-zA-Z0-9_]+)=(["\'])?(?P<value>(?(3)(?:(?:(?!\3|\\\\).|\\\\\\3)*)|(?:[^\s]+)))(?(3)\\3)/s', $str, $matches);
	$stripslashes = function (&$value, $key) {
		$value = stripslashes($value);
	};
	array_walk($matches['value'], $stripslashes);
	array_walk($matches[4], $stripslashes);
	$matches['machinetags'] = $matches[0];
	foreach($matches as $key=>$val) if (is_int($key)) unset($matches[$key]);
	return $matches;
}

print_r($test_cases);
print_r(extract_machine_tags($test_cases));


?>