mirror of
https://github.com/shlinkio/shlink.git
synced 2025-02-25 18:45:27 -06:00
Changed logic when resolving the title of a URL, to ensure only html content is tried to be downloaded, and only until the title tag has been parsed
This commit is contained in:
@@ -14,6 +14,9 @@ use Shlinkio\Shlink\Core\Options\UrlShortenerOptions;
|
||||
use Throwable;
|
||||
|
||||
use function preg_match;
|
||||
use function str_contains;
|
||||
use function str_starts_with;
|
||||
use function strtolower;
|
||||
use function trim;
|
||||
|
||||
use const Shlinkio\Shlink\TITLE_TAG_VALUE;
|
||||
@@ -56,8 +59,18 @@ class UrlValidator implements UrlValidatorInterface, RequestMethodInterface
|
||||
return null;
|
||||
}
|
||||
|
||||
$body = $response->getBody()->__toString();
|
||||
preg_match(TITLE_TAG_VALUE, $body, $matches);
|
||||
$contentType = strtolower($response->getHeaderLine('Content-Type'));
|
||||
if (! str_starts_with($contentType, 'text/html')) {
|
||||
return null;
|
||||
}
|
||||
|
||||
$collectedBody = '';
|
||||
$body = $response->getBody();
|
||||
// With streaming enabled, we can walk the body until the </title> tag is found, and then stop
|
||||
while (! str_contains($collectedBody, '</title>') && ! $body->eof()) {
|
||||
$collectedBody .= $body->read(1024);
|
||||
}
|
||||
preg_match(TITLE_TAG_VALUE, $collectedBody, $matches);
|
||||
return isset($matches[1]) ? trim($matches[1]) : null;
|
||||
}
|
||||
|
||||
@@ -73,6 +86,7 @@ class UrlValidator implements UrlValidatorInterface, RequestMethodInterface
|
||||
RequestOptions::IDN_CONVERSION => true,
|
||||
// Making the request with a browser's user agent makes the validation closer to a real user
|
||||
RequestOptions::HEADERS => ['User-Agent' => self::CHROME_USER_AGENT],
|
||||
RequestOptions::STREAM => true, // This ensures large files are not fully downloaded if not needed
|
||||
]);
|
||||
} catch (GuzzleException $e) {
|
||||
throw InvalidUrlException::fromUrl($url, $e);
|
||||
|
||||
@@ -132,11 +132,46 @@ class UrlValidatorTest extends TestCase
|
||||
$request->shouldHaveBeenCalledOnce();
|
||||
}
|
||||
|
||||
/** @test */
|
||||
public function validateUrlWithTitleReturnsNullWhenAutoResolutionIsEnabledAndReturnedContentTypeIsInvalid(): void
|
||||
{
|
||||
$request = $this->httpClient->request(RequestMethodInterface::METHOD_GET, Argument::cetera())->willReturn(
|
||||
new Response('php://memory', 200, ['Content-Type' => 'application/octet-stream']),
|
||||
);
|
||||
$this->options->autoResolveTitles = true;
|
||||
|
||||
$result = $this->urlValidator->validateUrlWithTitle('http://foobar.com/12345/hello?foo=bar', true);
|
||||
|
||||
self::assertNull($result);
|
||||
$request->shouldHaveBeenCalledOnce();
|
||||
}
|
||||
|
||||
/** @test */
|
||||
public function validateUrlWithTitleReturnsNullWhenAutoResolutionIsEnabledAndBodyDoesNotContainTitle(): void
|
||||
{
|
||||
$request = $this->httpClient->request(RequestMethodInterface::METHOD_GET, Argument::cetera())->willReturn(
|
||||
new Response($this->createStreamWithContent('<body>No title</body>'), 200, ['Content-Type' => 'text/html']),
|
||||
);
|
||||
$this->options->autoResolveTitles = true;
|
||||
|
||||
$result = $this->urlValidator->validateUrlWithTitle('http://foobar.com/12345/hello?foo=bar', true);
|
||||
|
||||
self::assertNull($result);
|
||||
$request->shouldHaveBeenCalledOnce();
|
||||
}
|
||||
|
||||
private function respWithTitle(): Response
|
||||
{
|
||||
$body = new Stream('php://temp', 'wr');
|
||||
$body->write('<title data-foo="bar"> Resolved title</title>');
|
||||
$body = $this->createStreamWithContent('<title data-foo="bar"> Resolved title</title>');
|
||||
return new Response($body, 200, ['Content-Type' => 'TEXT/html; charset=utf-8']);
|
||||
}
|
||||
|
||||
return new Response($body);
|
||||
private function createStreamWithContent(string $content): Stream
|
||||
{
|
||||
$body = new Stream('php://temp', 'wr');
|
||||
$body->write($content);
|
||||
$body->rewind();
|
||||
|
||||
return $body;
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user