XHP with Regex for link Replacement - php

Question

I am trying to implement a simple function that given a text input, returns the text modified with xhp_a when a link is detected, within a paragraph xhp_p.
Consider this class
class Urlifier {
protected static $reg_exUrl = "/(http|https|ftp|ftps)\:\/\/[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(\/\S*)?/";
public static function convertParagraphWithLink(?string $input):xhp_p{
if (!$input)
return <p></p>;
else
{
if (preg_match(self::$reg_exUrl,$input,$url_match)) //match found
{
return <p>{preg_replace($reg_exUrl, '<a href="'.$url_match[0].'>'.$url_match[0].'</a>', $input)}<p>;
}else{//no link inside
<p>{$input}</p>
}
}
}
The problem here is that xhp escapes html and links are not shown as expected. I suppose that this happens because a do not create a dom hierarchy as expected (with appendChild method for example) and thus everything regex replaces is a string.
So my other approach to this problem was to use preg_match_callback with a callback function that would create xhp_a and add to hierarchy under xhp_p but that did not work either.
Am i wrong somewhere ? If not would there by any security risk / bigger overhead by just finding and replacing on load the html on client side instead of server ?
Thanks for your time !

Answer 1

Since XHP maintains object hierarchy that maps to DOM, simply replacing parts of a string won't create any new objects. To manipulate XHP objects corresponding methods should be used, e.g. appendChild.
Here's an example of how what you need can be achieved with XHP manipulation.
class Urlifier {
public static function convertParagraphWithLink(
?string $input,
): xhp_p {
$url_pattern = re"/(http|https|ftp|ftps)\:\/\/[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(\/\S*)?/";
if (HH\Lib\Str\is_empty($input)) {
return <p/>;
}
$input = $input as nonnull;
// Extract links
$link_matches = HH\Lib\Regex\every_match($input, $url_pattern);
$links = HH\Lib\Vec\map($link_matches, $m ==> $m[0]);
$a_elements = HH\Lib\Vec\map($links, $link ==> <a href={$link}>{$link}</a>);
// Extract all pieces between matches
$texts = HH\Lib\Regex\split($input, $url_pattern);
$p_elements = HH\Lib\Vec\map($texts, $text ==> <p>{$text}</p>);
// Merge texts and links
$pairs = HH\Lib\Vec\zip($p_elements, $a_elements);
$elements = HH\Lib\Vec\flatten($pairs);
// Because there's one more p element than a element, append last p
$elements[] = HH\Lib\C\last($p_elements);
$result = <p/>;
$result->appendChild($elements);
return $result;
}