Crudely sniff the title of a web page reading as few bytes as possible

HttpTitleSniffer.php

<?php

namespace UFCOE;

/**
 * Sniff the title of a web page by reading as few bytes as possible.
 *
 * Warning, this is blissfully unaware of character encodings so that it can be fast. It only returns a string
 * if the markup appears ASCII compatible, but you could get back UTF-8, Windows-125*, ISO-8859-*, etc.
 */
class HttpTitleSniffer {

	public $httpContextOptions = array(
		'max_redirects' => 1,
		'timeout' => 5,
	);

	public $bytesPerRead = 1024;

	public function sniffTitle($url, $stream = null) {
		if (!$stream) {
			$stream = $this->getStream($url);
			if (!$stream) {
				return false;
			}
		}

		// read until it looks like we have the closing title tag
		$content = '';
		$hasTitle = false;
		while (!feof($stream) || !$hasTitle) {
			$content .= fread($stream, $this->bytesPerRead);
			$hasTitle = (false !== strpos($content, '</title>'));
		}
		fclose($stream);

		if (preg_match('~<title\b[^>]*>([^<]+)<~', $content, $m)) {
			return html_entity_decode($m[1], ENT_HTML401, 'UTF-8');
		}

		return false;
	}

	protected function getStream($url) {
		// see https://bugs.php.net/bug.php?id=51192
		$php_5_2_13_and_below = version_compare(PHP_VERSION, '5.2.14', '<');
		$php_5_3_0_to_5_3_2 = version_compare(PHP_VERSION, '5.3.0', '>=') &&
			version_compare(PHP_VERSION, '5.3.3', '<');

		if ($php_5_2_13_and_below || $php_5_3_0_to_5_3_2) {
			$tmpUrl = str_replace("-", "", $url);
			$validated = filter_var($tmpUrl, FILTER_VALIDATE_URL);
		} else {
			$validated = filter_var($url, FILTER_VALIDATE_URL);
		}
		if (!$validated) {
			return false;
		}

		// validator accepts all kinds of protocols
		if (!preg_match('~^https?\://~', $url)) {
			return false;
		}

		$ctx = stream_context_create(array('http' => $this->httpContextOptions));

		return fopen($url, 'r', false, $ctx);
	}
}