Ticket #3403: wxr-cats.diff

wp-admin/import/wordpress.php

@@ -85 +85 @@
                 $importdata = preg_replace("/(\r\n|\n|\r)/", "\n", $importdata);
                 preg_match_all('|<item>(.*?)</item>|is', $importdata, $this->posts);
                 $this->posts = $this->posts[1];
+                preg_match_all('|<wp:category>(.*?)</wp:category>|is', $importdata, $this->categories);
+                $this->categories = $this->categories[1];
         }
 
         function get_wp_authors() {
@@ -177 +179 @@
                 $this->wp_authors_form();
         }
 
+        function process_categories() {
+                global $wpdb;
+
+                $cat_names = (array) $wpdb->get_col("SELECT cat_name FROM $wpdb->categories");
+
+                while ( $c = array_shift($this->categories) ) {
+                        $cat_name = trim(str_replace(array ('<![CDATA[', ']]>'), '', $this->get_tag( $c, 'wp:cat_name' )));
+
+                        // If the category exists we leave it alone
+                        if ( in_array($cat_name, $cat_names) )
+                                continue;
+
+                        $category_nicename      = $this->get_tag( $c, 'wp:category_nicename' );
+                        $posts_private          = (int) $this->get_tag( $c, 'wp:posts_private' );
+                        $links_private          = (int) $this->get_tag( $c, 'wp:links_private' );
+
+                        $parent = $this->get_tag( $c, 'wp:category_parent' );
+
+                        if ( empty($parent) )
+                                $category_parent = '0';
+                        else
+                                $category_parent = (int) category_exists($parent);
+
+                        $catarr = compact('category_nicename', 'category_parent', 'posts_private', 'links_private', 'posts_private', 'cat_name');
+
+                        $cat_ID = wp_insert_category($catarr);
+                }
+        }
+
         function process_posts() {
                 global $wpdb;
                 $i = -1;
@@ -206 +237 @@
 
                         $cat_index = 0;
                         foreach ($categories as $category) {
-                                $categories[$cat_index] = $wpdb->escape($this->unhtmlentities($category));
+                                $categories[$cat_index] = $wpdb->escape($this->unhtmlentities(str_replace(array ('<![CDATA[', ']]>'), '', $category)));
                                 $cat_index++;
                         }
 
@@ -277 +308 @@
                 $this->file = get_attached_file($this->id);
                 $this->get_authors_from_post();
                 $this->get_entries();
+                $this->process_categories();
                 $this->process_posts();
         }
 

wp-admin/export.php

@@ -60 +60 @@
 }
 
 $posts = $wpdb->get_results("SELECT * FROM $wpdb->posts $where ORDER BY post_date_gmt ASC");
+
+$categories = (array) $wpdb->get_results("SELECT cat_ID, cat_name, category_nicename, category_description, category_parent, posts_private, links_private FROM $wpdb->categories LEFT JOIN $wpdb->post2cat ON (category_id = cat_id) LEFT JOIN $wpdb->posts ON (post_id <=> id) $where GROUP BY cat_id");
+
+function wxr_missing_parents($categories) {
+        if ( !is_array($categories) || empty($categories) )
+                return array();
+
+        foreach ( $categories as $category )
+                $parents[$category->cat_ID] = $category->category_parent;
+
+        $parents = array_unique(array_diff($parents, array_keys($parents)));
+
+        if ( $zero = array_search('0', $parents) )
+                unset($parents[$zero]);
+
+        return $parents;
+}
+
+while ( $parents = wxr_missing_parents($categories) ) {
+        $found_parents = $wpdb->get_results("SELECT cat_ID, cat_name, category_nicename, category_description, category_parent, posts_private, links_private FROM $wpdb->categories WHERE cat_ID IN (" . join(', ', $parents) . ")");
+        if ( is_array($found_parents) && count($found_parents) )
+                $categories = array_merge($categories, $found_parents);
+        else
+                break;
+}
+
+// Put them in order to be inserted with no child going before its parent
+$pass = 0;
+$passes = 1000 + count($categories);
+while ( ( $cat = array_shift($categories) ) && ++$pass < $passes ) {
+        if ( $cat->category_parent == 0 || isset($cats[$cat->category_parent]) ) {
+                $cats[$cat->cat_ID] = $cat;
+        } else {
+                $categories[] = $cat;
+        }
+}
+unset($categories);
+
+function wxr_cdata($str) {
+        if ( seems_utf8($str) == false )
+                $str = utf8_encode($str);
+
+        // $str = ent2ncr(wp_specialchars($str));
+
+        $str = "<![CDATA[$str" . ( ( substr($str, -1) == ']' ) ? ' ' : '') . "]]>";
+
+        return $str;
+}
+
+function wxr_cat_name($c) {
+        if ( empty($c->cat_name) )
+                return;
+
+        echo '<wp:cat_name>' . wxr_cdata($c->cat_name) . '</wp:cat_name>';
+}
+
+function wxr_category_description($c) {
+        if ( empty($c->category_description) )
+                return;
+
+        echo '<wp:category_description>' . wxr_cdata($c->category_description) . '</wp:category_description>';
+}
 ?>
 <!-- This is a WordPress eXtended RSS file generated by WordPress as an export of your blog. -->
 <!-- It contains information about your blog's posts, comments, and categories. -->
@@ -86 +148 @@
         <pubDate><?php echo mysql2date('D, d M Y H:i:s +0000', get_lastpostmodified('GMT'), false); ?></pubDate>
         <generator>http://wordpress.org/?v=<?php bloginfo_rss('version'); ?></generator>
         <language><?php echo get_option('rss_language'); ?></language>
+<?php if ( $cats ) : foreach ( $cats as $c ) : ?>
+        <wp:category><wp:category_nicename><?php echo $c->category_nicename; ?></wp:category_nicename><wp:category_parent><?php echo $c->category_parent ? $cats[$c->category_parent]->cat_name : ''; ?></wp:category_parent><wp:posts_private><?php echo $c->posts_private ? '1' : '0'; ?></wp:posts_private><wp:links_private><?php echo $c->links_private ? '1' : '0'; ?></wp:links_private><?php wxr_cat_name($c); ?><?php wxr_category_description($c); ?></wp:category>
+<?php endforeach; endif; ?>
         <?php do_action('rss2_head'); ?>
         <?php if ($posts) { foreach ($posts as $post) { start_wp(); ?>
 <item>