bookpathalias issue

Confirmed. On it and tracking it down.

OMG... Looking at our bookalias token code and I'm struggling to figure out why it's so complex.

We currently have:

          // Get the bookpathalias token
          $values['bookpathalias'] = '';
          // We're on a node and it's a book and it has a parent? Get the book path alias.
          if (module_exists('book') && !empty($object->book['menu_name'])) {
            for ($i = 1; $i < 10; $i++) {
              $second_child = $i + 2;
              if (!$object->book["p$second_child"]) {
                $parent_node = db_result(db_query('SELECT link_path FROM {menu_links} WHERE mlid = %d', $object->book["p$i"]));
                break;
              }
            $values['bookpathalias'] = drupal_get_path_alias($parent_node);
          }

When we *have* the parent information right in $node->book['plid'] (same thing book.module uses when creating its 'Up' navigation link on book pages).

      $values['bookpathalias'] = '';
      if (!empty($object->book['plid'])) {
        $parent_link = db_result(db_query("SELECT link_path FROM {menu_links} WHERE mlid = %d", $object->book['plid']));
        $values['bookpathalias'] = drupal_get_path_alias($parent_link);
      }

Is there a reason for the complex parent calculation?

And just to be clear, the following should be the expected behavior for this token so I can write a test for this:

Book node type pattern: [bookpathalias]/[node-title]

Node 1: First (root) page in book:
Title: Root page
Generated alias: root-page

Node 2: First sub-page of node 1
Title: First sub page
Generated alias: root-page/first-sub-page

Node 3: Second sub-page of node 1
Title Second sub page
Generated alias: root-page/second-sub-page

Node 4: First sub-sub page node 2
Title: First sub sub page
Generated alias: root-page/first-sub-page/first-sub-sub-page

Also, what should be done if a page doesn't have an alias? Should we return the un-aliased path (e.g. 'node/2') as the bookpathalias token?

#2 - I'm not really sure how we missed that aside from "plid" not being all that intuitive.

#3 and #4 - those seem right to me.

Here's the patch for 7.x-1.x with the very simplified parent fetching to run through the testbot. This would be backported to both 6.x branches.

Committed to all three branches + passing tests.
http://drupal.org/cvs?commit=331066
http://drupal.org/cvs?commit=331068
http://drupal.org/cvs?commit=331070
http://drupal.org/cvs?commit=331072
http://drupal.org/cvs?commit=331076
http://drupal.org/cvs?commit=331078