tec: Handle feeds declaring a wrong encoding

flusio · Sep 9, 2022 · 0cb9840 · 0cb9840
1 parent fe2e227
commit 0cb9840
Show file tree

Hide file tree

Showing 2 changed files with 42 additions and 1 deletion.
diff --git a/lib/SpiderBits/src/feeds/Feed.php b/lib/SpiderBits/src/feeds/Feed.php
@@ -50,7 +50,26 @@ public static function fromText($feed_as_string)
         $dom_document = new \DOMDocument();
         $result = @$dom_document->loadXML($feed_as_string);
         if (!$result) {
-            throw new \DomainException('Can’t parse the given string.');
+            // It might be an encoding issue. We try to recover by re-encoding
+            // the string with the declared encoding, or UTF-8. It will most
+            // probably generate a string with characters replaced by `?`, but
+            // at least it will be parsable.
+            $result = preg_match(
+                '/<?xml\s+(?:(?:.*?)\s)?encoding="(.+?)"/i',
+                $feed_as_string,
+                $matches
+            );
+            if ($result) {
+                $encoding = $matches[1];
+            } else {
+                $encoding = 'UTF-8';
+            }
+            $feed_as_string = mb_convert_encoding($feed_as_string, $encoding, $encoding);
+            $result = @$dom_document->loadXML($feed_as_string);
+
+            if (!$result) {
+                throw new \DomainException('Can’t parse the given string.');
+            }
         }
 
         if (AtomParser::canHandle($dom_document)) {

diff --git a/tests/lib/SpiderBits/feeds/FeedTest.php b/tests/lib/SpiderBits/feeds/FeedTest.php
@@ -238,6 +238,28 @@ public function testFromTextWithEmptyRss()
         $this->assertSame('4a022608d595e9000d1f1be22a0a6a0763ad853d2417b1c8ea0ea12bd047bdcd', $feed->hash());
     }
 
+    public function testFromTextRecoversFromWrongEncoding()
+    {
+        // Create a XML string declaring encoding UTF-8
+        $xml = <<<XML
+            <?xml version="1.0" encoding="UTF-8"?>
+            <rss version="2.0">
+                <channel>
+                    <title>My feed with an àccent</title>
+                    <link>https://example.com</link>
+                </channel>
+            </rss>
+            XML;
+        // … but its real encoding is ISO-8859-1!
+        $xml = mb_convert_encoding($xml, 'ISO-8859-1', 'UTF-8');
+
+        $feed = Feed::fromText($xml);
+
+        $this->assertSame('My feed with an ?ccent', $feed->title);
+        $this->assertSame('https://example.com', $feed->link);
+        $this->assertSame('0ca0efc2ed6d8bfab90dbd3f42a473465c505b3d13b4da5975b0deed52c4b231', $feed->hash());
+    }
+
     public function testFromTextFailsWithEmptyString()
     {
         $this->expectException(\DomainException::class);