Added charset detection from HTTP headers. Fixes #1

2 files changed, 71 insertions, 0 deletions

diff --git a/tests/charset.php b/tests/charset.php
new file mode 100644
index 0000000..f429241
--- /dev/null
+++ b/tests/charset.php
@@ -0,0 +1,60 @@
+<?php
+
+$config = array(
+    'type' => 'xpath',
+    'xpath' => 'div[@class="meldung_wrapper"]',
+);
+
+// http://www.heise.de/newsticker/heise-atom.xml
+
+$article = array(
+    'link' => 'http://www.heise.de/newsticker/meldung/Fruehjahrspatches-Microsoft-9-Adobe-3-1838175.html/from/atom10',
+    'content' => 'This is the feed content',
+    'plugin_data' => '',
+);
+
+$doc = new DOMDocument();
+
+$html = file_get_contents($article['link']);
+
+// BEGIN --- New code
+$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
+    }
+}
+
+$charset = false;
+if ($content_type) {
+    preg_match('/charset=(\S+)/', $content_type, $matches);
+    if (isset($matches[1]) && !empty($matches[1])) $charset = $matches[1];
+}
+
+// END --- New code
+
+echo 'CHARSET: ' . $charset . PHP_EOL;
+
+
+$doc->loadHTML('<?xml encoding="' . $charset . '">' . $html);
+
+echo 'ENCODING: ' . $doc->encoding . PHP_EOL;
+
+if ($doc) {
+    $basenode = false;
+    $xpath = new DOMXPath($doc);
+    $entries = $xpath->query('(//'.$config['xpath'].')');   // find main DIV according to config
+
+    var_dump($entries);
+
+    if ($entries->length > 0) $basenode = $entries->item(0);
+
+    if ($basenode) {
+       $article['content'] = $doc->saveXML($basenode);
+       $article['plugin_data'] = "feedmod,$owner_uid:" . $article['plugin_data'];
+    }
+}
+
+print_r($article);
diff --git a/tests/charsetcurl.php b/tests/charsetcurl.php
new file mode 100644
index 0000000..9db93e8
--- /dev/null
+++ b/tests/charsetcurl.php
@@ -0,0 +1,11 @@
+<?php
+
+$ch = curl_init('http://www.heise.de/newsticker/meldung/Fruehjahrspatches-Microsoft-9-Adobe-3-1838175.html/from/atom10');
+#curl_setopt($ch, CURLOPT_URL, "http://www.example.com/");
+#curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+#curl_setopt($ch, CURLINFO_HEADER_OUT, true);
+
+curl_exec($ch);
+
+var_dump(curl_getinfo($ch,CURLINFO_CONTENT_TYPE));
+