[php-blog-cvs] serendipity/include/admin/importers generic.inc.php,1.5,1.6

[Garvin H. <gar...@us...>]

Update of /cvsroot/php-blog/serendipity/include/admin/importers
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv4943/include/admin/importers

Modified Files:
	generic.inc.php 
Log Message:
If RSS feeds are encoded differently than UTF-8, we had a problem.
Let users specify the source encoding of an RSS feed, because Onyx RSS doesn't report the used charset.

Fixes import problems and displaying remote RSS feeds.


Index: generic.inc.php
===================================================================
RCS file: /cvsroot/php-blog/serendipity/include/admin/importers/generic.inc.php,v
retrieving revision 1.5
retrieving revision 1.6
diff -u -d -r1.5 -r1.6
--- generic.inc.php	26 Jan 2005 14:14:05 -0000	1.5
+++ generic.inc.php	15 Feb 2005 11:25:26 -0000	1.6
@@ -10,6 +10,15 @@
     var $inputFields = array();
 
     function Serendipity_Import_Generic($data) {
+        $charsets = array();
+        if (LANG_CHARSET != 'UTF-8') {
+            $charsets['UTF-8'] = 'UTF-8';
+        }
+        if (LANG_CHARSET != 'ISO-8859-1') {
+            $charset['ISO-8859-1'] = 'ISO-8859-1';
+        }
+        $charsets['native'] = LANG_CHARSET;
+
         $this->data = $data;
         $this->inputFields = array(array('text'    => RSS . ' ' . URL,
                                          'type'    => 'input',
@@ -26,6 +35,12 @@
                                          'name'    => 'category',
                                          'value'   => 0,
                                          'default' => $this->_getCategoryList()),
+                                         
+                                   array('text'    => CHARSET,
+                                         'type'    => 'list',
+                                         'name'    => 'charset',
+                                         'value'   => 'UTF-8',
+                                         'default' => $charsets),
 
                                     array('text'   => RSS_IMPORT_BODYONLY,
                                          'type'    => 'bool',
@@ -52,6 +67,26 @@
         return $ret;
     }
 
+    function decode($string) {
+        switch($this->data['charset']) {
+            case 'native':
+                return $string;
+
+            case 'ISO-8859-1':
+                if (function_exists('iconv')) {
+                    return iconv('ISO-8859-1', LANG_CHARSET, $string);
+                } elseif (function_exists('recode')) {
+                    return recode('iso-8859-1..' . LANG_CHARSET, $string);
+                } else {
+                    return $string;
+                }
+            
+            case 'UTF-8':
+            default:
+                return utf8_decode($string);
+        }
+    }
+
     function buildEntry($item, &$entry) {
         global $serendipity;
 
@@ -59,7 +94,7 @@
         $bodyonly = serendipity_get_bool($this->data['bodyonly']);
 
         if ($item['description']) {
-            $entry['body'] = utf8_decode($item['description']);
+            $entry['body'] = $this->decode($item['description']);
         }
 
         if ($item['content:encoded']) {
@@ -75,14 +110,14 @@
             // switch). We substract 4 letters because of possible '...' additions to an entry.
             $testbody = substr(trim(strip_tags($entry['body'])), 0, -4);
             if ($testbody != substr(trim(strip_tags($item['content:encoded'])), 0, strlen($testbody))) {
-                $data .= utf8_decode($item['content:encoded']);
+                $data .= $this->decode($item['content:encoded']);
             } else {
-                $data = utf8_decode($item['content:encoded']);
+                $data = $this->decode($item['content:encoded']);
             }
         }
 
-        $entry['title'] = utf8_decode($item['title']);
-        $entry['timestamp'] = utf8_decode(strtotime(isset($item['pubdate']) ? $item['pubdate'] : $item['dc:date']));
+        $entry['title'] = $this->decode($item['title']);
+        $entry['timestamp'] = $this->decode(strtotime(isset($item['pubdate']) ? $item['pubdate'] : $item['dc:date']));
         if ($entry['timestamp'] == -1) {
             $entry['timestamp'] = time();
         }
@@ -94,7 +129,7 @@
         }
 
         if (!empty($item['category'])) {
-            $cat = serendipity_fetchCategoryInfo(0, trim(utf8_decode($item['category'])));
+            $cat = serendipity_fetchCategoryInfo(0, trim($this->decode($item['category'])));
             if (is_array($cat) && isset($cat['categoryid'])) {
                 $entry['categories'][] = $cat['categoryid'];
             }