Added charset detection from HTTP headers. Fixes #1

1 files changed, 37 insertions, 1 deletions

diff --git a/init.php b/init.php
index e4b93fe..ddfde29 100644
--- a/init.php
+++ b/init.php
@@ -46,6 +46,8 @@ class Af_Feedmod extends Plugin implements IHandler
 
     function hook_article_filter($article)
     {
+        global $fetch_last_content_type;
+
         $json_conf = $this->host->get($this, 'json_conf');
         $owner_uid = $article['owner_uid'];
         $data = json_decode($json_conf, true);
@@ -61,7 +63,41 @@ class Af_Feedmod extends Plugin implements IHandler
             switch ($config['type']) {
                 case 'xpath':
                     $doc = new DOMDocument();
-                    @$doc->loadHTML(fetch_file_contents($article['link']));
+
+                    if (version_compare(VERSION, '1.7.9', '>=')) {
+                        $html = fetch_file_contents($article['link']);
+                        $content_type = $fetch_last_content_type;
+                    } else {
+                        // fallback to file_get_contents()
+                        $html = file_get_contents($article['link']);
+
+                        // try to fetch charset from HTTP headers
+                        $headers = $http_response_header;
+                        $content_type = false;
+                        foreach ($headers as $h) {
+                            if (substr(strtolower($h), 0, 13) == 'content-type:') {
+                                $content_type = substr($h, 14);
+                                // don't break here to find LATEST (if redirected) entry
+                            }
+                        }
+                    }
+
+                    if (!isset($config['force_charset'])) {
+                        $charset = false;
+                        if ($content_type) {
+                            preg_match('/charset=(\S+)/', $content_type, $matches);
+                            if (isset($matches[1]) && !empty($matches[1])) $charset = $matches[1];
+                        }
+
+                        if ($charset) {
+                            $html = '<?xml encoding="' . $charset . '">' . $html;
+                        }
+                    } else {
+                        // use forced charset
+                        $html = '<?xml encoding="' . $config['force_charset'] . '">' . $html;
+                    }
+
+                    @$doc->loadHTML($html);
 
                     if ($doc) {
                         $basenode = false;