clones/friendica
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Merge pull request #12513 from MrPetovan/bug/12507-rel-meeee

Browse Source 
Add call to HTML::checkRelMeLink in Worker\CheckRelMeProfileLink
This commit is contained in:
Philipp
2022-12-23 21:12:06 +01:00
committed by GitHub
parent e4ac3c7922 b2f2fbc4d4
commit 8fd0d4cdc0
10 changed files with 303 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
src/Content/Text/HTML.php
View File

@@ -33,6 +33,7 @@ use Friendica\Util\Network;
use Friendica\Util\Strings;
use Friendica\Util\XML;
use League\HTMLToMarkdown\HtmlConverter;
use Psr\Http\Message\UriInterface;
class HTML
{
@@ -1007,4 +1008,51 @@ class HTML
return $text;
}
/**
* XPath arbitrary string quoting
*
* @see https://stackoverflow.com/a/45228168
* @param string $value
* @return string
*/
public static function xpathQuote(string $value): string
{
if (false === strpos($value, '"')) {
return '"' . $value . '"';
}
if (false === strpos($value, "'")) {
return "'" . $value . "'";
}
// if the value contains both single and double quotes, construct an
// expression that concatenates all non-double-quote substrings with
// the quotes, e.g.:
//
// concat("'foo'", '"', "bar")
return 'concat(' . implode(', \'"\', ', array_map(['self', 'xpathQuote'], explode('"', $value))) . ')';
}
/**
* Checks if the provided URL is present in the DOM document in an element with the rel="me" attribute
*
* XHTML Friends Network http://gmpg.org/xfn/
*
* @param DOMDocument $doc
* @param UriInterface $meUrl
* @return bool
*/
public static function checkRelMeLink(DOMDocument $doc, UriInterface $meUrl): bool
{
$xpath = new \DOMXpath($doc);
// This expression checks that "me" is among the space-delimited values of the "rel" attribute.
// And that the href attribute contains exactly the provided URL
$expression = "//*[contains(concat(' ', normalize-space(@rel), ' '), ' me ')][@href = " . self::xpathQuote($meUrl) . "]";
$result = $xpath->query($expression);
return $result !== false && $result->length > 0;
}
}

72
src/Worker/CheckRelMeProfileLink.php
View File

@@ -22,14 +22,14 @@
namespace Friendica\Worker;
use DOMDocument;
use Friendica\DI;
use Friendica\Content\Text\HTML;
use Friendica\Core\Logger;
use Friendica\DI;
use Friendica\Model\Profile;
use Friendica\Model\User;
use Friendica\Network\HTTPClient\Client\HttpClientAccept;
use Friendica\Network\HTTPClient\Client\HttpClientOptions;
use Friendica\Util\Network;
use Friendica\Util\Strings;
use GuzzleHttp\Psr7\Uri;
/* This class is used to verify the homepage link of a user profile.
* To do so, we look for rel="me" links in the given homepage, if one
@@ -56,43 +56,37 @@ class CheckRelMeProfileLink
{
Logger::notice('Verifying the homepage', ['uid' => $uid]);
Profile::update(['homepage_verified' => false], $uid);
$homepageUrlVerified = false;
$owner = User::getOwnerDataById($uid);
if (!empty($owner['homepage'])) {
$xrd_timeout = DI::config()->get('system', 'xrd_timeout');
$curlResult = DI::httpClient()->get($owner['homepage'], $accept_content = HttpClientAccept::HTML, [HttpClientOptions::TIMEOUT => $xrd_timeout]);
if ($curlResult->isSuccess()) {
$content = $curlResult->getBody();
if (!$content) {
Logger::notice('Empty body of the fetched homepage link). Cannot verify the relation to profile of UID %s.', ['uid' => $uid, 'owner homepage' => $owner['homepage']]);
} else {
$doc = new DOMDocument();
@$doc->loadHTML($content);
if (!$doc) {
Logger::notice('Could not parse the content');
} else {
foreach ($doc->getElementsByTagName('a') as $link) {
$rel = $link->getAttribute('rel');
if ($rel == 'me') {
$href = $link->getAttribute('href');
if (!$homepageUrlVerified && Network::isValidHttpUrl($href)) {
$homepageUrlVerified = Strings::compareLink($owner['url'], $href);
}
}
}
}
if ($homepageUrlVerified) {
Profile::update(['homepage_verified' => true], $uid);
Logger::notice('Homepage URL verified', ['uid' => $uid, 'owner homepage' => $owner['homepage']]);
} else {
Logger::notice('Homepage URL could not be verified', ['uid' => $uid, 'owner homepage' => $owner['homepage']]);
}
}
} else {
Logger::notice('Could not cURL the homepage URL', ['owner homepage' => $owner['homepage']]);
}
} else {
$owner = User::getOwnerDataById($uid);
if (empty($owner['homepage'])) {
Logger::notice('The user has no homepage link.', ['uid' => $uid]);
return;
}
$xrd_timeout = DI::config()->get('system', 'xrd_timeout');
$curlResult = DI::httpClient()->get($owner['homepage'], HttpClientAccept::HTML, [HttpClientOptions::TIMEOUT => $xrd_timeout]);
if (!$curlResult->isSuccess()) {
Logger::notice('Could not cURL the homepage URL', ['owner homepage' => $owner['homepage']]);
return;
}
$content = $curlResult->getBody();
if (!$content) {
Logger::notice('Empty body of the fetched homepage link). Cannot verify the relation to profile of UID %s.', ['uid' => $uid, 'owner homepage' => $owner['homepage']]);
return;
}
$doc = new DOMDocument();
if (!@$doc->loadHTML($content)) {
Logger::notice('Could not parse the content');
return;
}
if (HTML::checkRelMeLink($doc, new Uri($owner['url']))) {
Profile::update(['homepage_verified' => true], $uid);
Logger::notice('Homepage URL verified', ['uid' => $uid, 'owner homepage' => $owner['homepage']]);
} else {
Logger::notice('Homepage URL could not be verified', ['uid' => $uid, 'owner homepage' => $owner['homepage']]);
}
}
}