Handle redirects when scraping feed from HTML (#7654)

* Handle redirects when scraping feed from HTML

* pass codesniffer

* pass PHPStan

* Optimize

* Another approach relying on HTML base
Standard way to save an HTML document with relative references

* Fix case of existing HTML base
which should not be overriden

---------

Co-authored-by: Alexandre Alapetite <alexandre@alapetite.fr>

3 files changed, 10 insertions, 6 deletions

diff --git a/app/Models/Category.php b/app/Models/Category.php
index e883a99cf..554e002fb 100644
--- a/app/Models/Category.php
+++ b/app/Models/Category.php
@@ -188,7 +188,7 @@ class FreshRSS_Category extends Minz_Model {
 		}
 		$ok = true;
 		$cachePath = $this->cacheFilename($url);
-		$opml = httpGet($url, $cachePath, 'opml', $this->attributes(), $this->curlOptions());
+		$opml = httpGet($url, $cachePath, 'opml', $this->attributes(), $this->curlOptions())['body'];
 		if ($opml == '') {
 			Minz_Log::warning('Error getting dynamic OPML for category ' . $this->id() . '! ' .
 				\SimplePie\Misc::url_remove_credentials($url));
diff --git a/app/Models/Entry.php b/app/Models/Entry.php
index bc5ed2279..66c05a830 100644
--- a/app/Models/Entry.php
+++ b/app/Models/Entry.php
@@ -845,7 +845,7 @@ HTML;
 	 * @param string $url Overridden URL. Will default to the entry URL.
 	 * @throws Minz_Exception
 	 */
-	public function getContentByParsing(string $url = '', int $maxRedirs = 3): string {
+	public function getContentByParsing(string $url = '', int $maxRedirs = 4): string {
 		$url = $url ?: htmlspecialchars_decode($this->link(), ENT_QUOTES);
 		$feed = $this->feed();
 		if ($url === '' || $feed === null || $feed->pathEntries() === '') {
@@ -872,12 +872,16 @@ HTML;
 		}
 
 		$cachePath = $feed->cacheFilename($url . '#' . $feed->pathEntries());
-		$html = httpGet($url, $cachePath, 'html', $feed->attributes(), $feed->curlOptions());
-		if (strlen($html) > 0) {
+		$response = httpGet($url, $cachePath, 'html', $feed->attributes(), $feed->curlOptions());
+		$html = $response['body'];
+		if ($html !== '') {
 			$doc = new DOMDocument();
 			$doc->loadHTML($html, LIBXML_NONET | LIBXML_NOERROR | LIBXML_NOWARNING);
 			$xpath = new DOMXPath($doc);
 
+			// Account for HTTP redirections
+			$url = $response['effective_url'] ?: $url;
+			$maxRedirs -= $response['redirect_count'];
 			if ($maxRedirs > 0) {
 				//Follow any HTML redirection
 				$metas = $xpath->query('//meta[@content]') ?: [];
diff --git a/app/Models/Feed.php b/app/Models/Feed.php
index fc17c875f..3c5fed507 100644
--- a/app/Models/Feed.php
+++ b/app/Models/Feed.php
@@ -788,7 +788,7 @@ class FreshRSS_Feed extends Minz_Model {
 		}
 
 		$httpAccept = $this->kind() === FreshRSS_Feed::KIND_HTML_XPATH_JSON_DOTNOTATION ? 'html' : 'json';
-		$content = httpGet($feedSourceUrl, $this->cacheFilename(), $httpAccept, $this->attributes(), $this->curlOptions());
+		$content = httpGet($feedSourceUrl, $this->cacheFilename(), $httpAccept, $this->attributes(), $this->curlOptions())['body'];
 		if (strlen($content) <= 0) {
 			return null;
 		}
@@ -846,7 +846,7 @@ class FreshRSS_Feed extends Minz_Model {
 		}
 
 		$httpAccept = $this->kind() === FreshRSS_Feed::KIND_XML_XPATH ? 'xml' : 'html';
-		$html = httpGet($feedSourceUrl, $this->cacheFilename(), $httpAccept, $this->attributes(), $this->curlOptions());
+		$html = httpGet($feedSourceUrl, $this->cacheFilename(), $httpAccept, $this->attributes(), $this->curlOptions())['body'];
 		if (strlen($html) <= 0) {
 			return null;
 		}