Import terms in the absence of <wp:category>, <wp:tag> data

It seems that a lot of people learn the hard way (like I did) that if you export anything but "All content" from WordPress you end up with a file that is missing a lot of essential stuff, including the wp:tag and wp:category tags that define these terms.

There was a thread on another issue which seemed to indicate that functionality was added that would trigger the import of terms attached to items even if the wp:tag and wp:category tags are not included in the file.

This is the issue/comment as well as the related commit:

Issue #1565032: Categories or Tags not imported
Commit: Issue #1565032 by mikeryan: Create terms on-the-fly if necessary

I haven't fully examined the code, but reading the thread it sounds like that's what it's supposed to do. If this is correct, then I'm afraid something may have gone wrong along the way as it no longer seems to work.

A quick look at wordpress_item.inc in most recent -dev (paste below) makes me wonder if the tracking vars $tagField and $categoryField will always be NULL, and perhaps this is the source of the problem?

class WordPressItemSource extends MigrateSourceXML {
  
  // ... a bunch of code omitted ...

  /**
   * Track the fields for the term references, so we can fix up if necessary
   * in prepare().
   *
   * @var string
   */
  protected $tagField = NULL;
  protected $categoryField = NULL;


  /**
   * Set it up
   */
  public function __construct(array $arguments = array()) {
    
    // ...

    if (module_exists('taxonomy')) {
      // Map the source fields to the configured vocabularies
      if ($arguments['tag_field']) {
        $this->addFieldMapping($arguments['tag_field'], 'tag')
          ->sourceMigration($arguments['group_name'] . $arguments['tag_migration'])
          ->xpath('category[@domain="post_tag"]/@nicename');
        $this->addFieldMapping($arguments['tag_field'] . ':source_type')
             ->defaultValue('tid');
      }
      else {
        $this->addFieldMapping(NULL, 'tag');
      }
      if ($arguments['category_field']) {
        $this->addFieldMapping($arguments['category_field'], 'category')
          ->sourceMigration($arguments['group_name'] . $arguments['category_migration'])
          ->xpath('category[@domain="category"]/@nicename');
        $this->addFieldMapping($arguments['category_field'] . ':source_type')
             ->defaultValue('tid');
      }
      else {
        $this->addFieldMapping(NULL, 'category');
      }
    }

    // ... 

  }

  // ...

  public function prepare(stdClass $node, stdClass $row) {
    
    // .. more code omitted ..

    // If any term relationships were unresolved, create them the hard way
    foreach (array('tag', 'category') as $term_type) {
      $meta_field_name = $term_type . 'Field';
      if ($this->$meta_field_name) {
        $field_name = $this->$meta_field_name;
        $value_name = $term_type . '_value';
        // Shortcut - if the counts match, don't need to dig deeper
        $field_values = field_get_items('node', $node, $field_name);
        if (!empty($field_values)) {
          $node_count = count($field_values);
        }
        else {
          $node_count = 0;
        }
        if (isset($row->$term_type)) {
          $row_count = count($row->$term_type);
        }
        else {
          $row_count = 0;
        }
        if ($node_count != $row_count) {
          $vid_field = $term_type . 'Vid';
          $vid = $this->$vid_field;
          // Get any terms already in the field
          $done_terms = array();
          if (is_array($field_values)) {
            foreach ($field_values as $value_array) {
              $terms = taxonomy_term_load_multiple($value_array);
              foreach ($terms as $term) {
                $done_terms[] = $term->name;
              }
            }
          }
          if (isset($row->$value_name)) {
            $diff = array_diff($row->$value_name, $done_terms);
            $field_language = field_language('node', $node, $field_name);
            foreach ($diff as $new_term_name) {
              // Let's see if the term already exists
              $matches = taxonomy_term_load_multiple(array(),
                array('name' => trim($new_term_name), 'vid' => $vid));
              if ($matches) {
                $node->{$field_name}[$field_language][] = array('tid' => key($matches));
              }
              else {
                $term = new stdClass;
                $term->name = $new_term_name;
                $term->vid = $vid;
                taxonomy_term_save($term);
                $node->{$field_name}[$field_language][] = array('tid' => $term->tid);
              }
            }
          }
        }
      }
    }
  }

  // ...

}

If I'm wrong about the functionality this provides, then please change this to a feature request. It would be excellent to have.

Thanks!