Merge pull request #48 from FreshRSS/fix-regex-greedy

Alkarex · web-flow · commit f3c4dbf34173 · 2025-08-11T00:35:59.000+02:00
Fix regex backtrack limit in clean_hash()
diff --git a/src/SimplePie.php b/src/SimplePie.php
@@ -1695,7 +1695,7 @@ private function clean_hash(string $rss): string
                         [
                             '#<(lastBuildDate|pubDate|updated|feedDate|dc:date|slash:comments)>[^<]+</\\1>#',
                             '#<(media:starRating|media:statistics) [^/<>]+/>#',
-                            '#<!--.+?-->#s',
+                            '#<!--.{,8192}?(-->|(?=]]>))#s', // XML comments up to a max length and stops at apparent end of CDATA section
                         ],
                         '',
                         $stream_data

Original file line number	Diff line number	Diff line change
`@@ -1695,7 +1695,7 @@ private function clean_hash(string $rss): string`
`1695`	`1695`	`[`
`1696`	`1696`	`'#<(lastBuildDate\|pubDate\|updated\|feedDate\|dc:date\|slash:comments)>[^<]+</\\1>#',`
`1697`	`1697`	`'#<(media:starRating\|media:statistics) [^/<>]+/>#',`
`1698`		`- '#<!--.+?-->#s',`
	`1698`	`+ '#<!--.{,8192}?(-->\|(?=]]>))#s', // XML comments up to a max length and stops at apparent end of CDATA section`
`1699`	`1699`	`],`
`1700`	`1700`	`'',`
`1701`	`1701`	`$stream_data`