Problems with Router / URL Matcher class - php

Question

since regular expressions aren't really my specialty, I need help with this little problem (in PHP).
I want to match a given url with an array of defined routes, e.g.:
$definedRoute = '/admin/user/[:id]/edit';
$url = '/admin/user/37/edit';
In my class, it would be like this, I imagine (getRoutes() returns an array of defined routes):
foreach ($this->getRoutes() as $route) {
$pattern = '~' . preg_replace('~\[\:[a-z]+\]~', '[a-z0-9]+',
str_replace('/', '\/', $route['definition'])) . '~';
if (preg_match($pattern, $url)) {
$parameters = $this->getRouteParameters($route['definition']);
(new $route['class']())->{$route['method']}($parameters);
// die? break?
}
}
I went about it like this: replace every occurence of a named parameter like [:id] with a regex for lowercase letters and numbers, e.g. [a-z0-9]+.
This would actually work but in some cases, it would match multiple and therefore the wrong routes. Also, it would always match ~\/~ in most cases. But every url should only be matched once.
Edit #1: the problem is: routes get matched multiple times. How can I prevent this?
Can someone enlighten me?

Answer 1

I don't know if this will cover every conceivable case, but you can use preg_match_all or preg_match as opposed to iterating over the patterns. It should also improve performance.
What this does is make the match order (left to right) important, with an array and loop you cannot do that (you can actually but it's uglier). Then we can sort the routes on complexity, like this:
//this is intentionally in the opposite order of what I want it.
$routes = ['definition' => ['\/', '\/admin\/user\/[a-z0-9]+\/']];
//the more / separators the closer to the beginning we want it. or the more complex regexs go first.
uasort($routes['definition'], function($a,$b){
//count the number of / in the route
//note the <=> spaceship (as it's called) is only available in PHP7+
return substr_count($b, '/') <=> substr_count($a, '/');
});
$url = '/admin/user/37/edit';
//in regex the pipe | is OR
preg_match('~^('.implode('|', $routes['definition']).')~i', $url, $matches);
print_r($matches);
Sandbox
Output:
Array
(
[0] => Array
(
[0] => /admin/user/37/edit
)
[1] => Array
(
[0] => /admin/user/37/edit
)
)
Which is correct in this case. Even if you do get multiple matches, you can count the length of them strlen and then take the longest or "best" match from them. This is pretty simple using strlen and probably sorting by length, so I will leave it up to you.
However I wouldn't call this a guarantee of it working 100% of the time, it's just the first thing that came to me.
Another Idea
Another idea is you are not anchoring the match to the start and end of the string. In theory the route could/should match the entire string so with my above example if you add ^ and $ here:
preg_match('~^('.implode('|', $routes['definition']).')$~i', $url, $matches);
This will ensure a full match and in this case ~\/~ will not match even if the array is not sorted, as you can see in the sandbox below.
Sandbox
That said it's not inconceivable you would only have/need a partial match. This is up to you and how you build your routes and URLs. You can of course just use the ^ start as with a begins with type of match, but you would need to sort them in that case.
Preg Match vs Preg Match All
Preg match will also work, but it will only return the first match. So if it matches multiple time you cannot compare them to find the best one. This may be fine if you use ^ and $.
Hope it helps.