Calling node_view within a filter can cause an infinite loop

Note: I tested this in Drupal 8.x not 7.x

I was not able to reproduce the recursion issue. The issue summary was a little vague, so it's possible, I'm not even doing what the original reporter was.

Here are my steps:

• Install the latest Drupal 8.x using the standard profile.
• Create a new node with title "node 1" and body empty
• Create a new node with title "node 2" and body "node 2 content"
• Enable php module
• Edit node1 with the following contents:

<?php
$node2 = node_load(2);
echo drupal_render(node_view($node2));
?>

I just ran the same steps on D7. It worked as expected, no errors.

I'm changing status to postponed until we get more info from the original reporter.

I really need some sample code and detailed steps to reproduce to be able to understand what you are trying to do. Please see my comment #2 for an example of what steps to reproduce should look like.

Setting status back for more detailed steps to reproduce this.

Similar problem here.

Trying to write a filter to embed nodes I get an infinite Loop after saving content.
The node's content contains a tag [task]%nid[/task] wich needs to be replaced by the renderend node output.

Strange thing: After calling any other page everything works fine and the rendered content appears as expected.

Here is my custom filter code:

function section_content_filter_info() {
  $filters['filter_content_tasks'] = array(
    'title' => t('Content - Tasks Filter'),
    'description' => t('Every [task](node_id)[/task] pattern is beeing replaced by its content.'),
    'process callback' => '_section_content_filter_tasks_process',
  );
  return $filters;
}

function _section_content_filter_tasks_process($text, $filter, $format) {
  $regex = '#\[task\]([0-9]+)\[\/task\]#s';
  preg_match($regex, $text, $matches);
  if (!empty($matches) && count($matches) > 1) {
    $nid = $matches[1];
    $node = node_load($nid);

    if ($node != FALSE && node_access('view', $node) && $node->status) {
      $node_view = node_view($node); //infinite loop here
      $node_output = drupal_render($node_view);
      $text = preg_replace($regex, $node_output, $text);
    }
  }
  return $text;
}

A non-existent viewname or other viewname than 'full' solves this issue for me.

$node_view = node_view($node, 'teaser', NULL);

@nm-dev: In your filter's info declaration, add 'cache' => FALSE. This fixed it for me (after much consternation!). E.g.,

function section_content_filter_info() {
  $filters['filter_content_tasks'] = array(
    'title' => t('Content - Tasks Filter'),
    'description' => t('Every [task](node_id)[/task] pattern is beeing replaced by its content.'),
    'process callback' => '_section_content_filter_tasks_process',

    'cache' => FALSE,

  );
  return $filters;
}

I can confirm the problem in the original post and #7. You must create a text filter that will load/view a node as part of its process (as seen in #7 code example).

I received the error Fatal error: Maximum function nesting level of '100' reached, aborting! in \includes\database\database.inc on line 248

I believe the workflow causing the error is this (not investigated, just a hunch):
1: You view node/1, which has a link in its body to node/2.
2: The filter acting on node/1 body loads node/2.
3: Whilst drupal is loading node/2, node/2's body needs text filters applied, which in turn may load node/3.
4: And the cycle continues...

function my_module_filter_process($text, $filter) {

  if (!preg_match_all('|\[\[node:(\d+)]]|', $text, $matches, PREG_SET_ORDER)) {
    return $text;
  }

  $nids = array();
  foreach ($matches as $match) {
    $nids[] = $match[1];
  }

  $query = db_select('field_data_field_category','f')
    ->fields('n',array('nid','title'))
    ->fields('t',array('tid','name'))
    ->condition('f.entity_id',$nids,'IN');
  $query->join('taxonomy_term_data','t','f.field_category_target_id = t.tid');
  $query->join('node','n','f.entity_id = n.nid');

  $results = $query->execute()->fetchAllAssoc('nid');
  
  foreach ($matches as $match) {
    $link = l("{$results[$match[1]]->name}: {$results[$match[1]]->title}", "node/{$match[1]}", array('attributes' => array('class' => 'inter-wiki-link')));
    $text = str_replace($match[0], $link, $text);
  }
  
  return $text;

}

I've just encountered this very issue, the problem is I can't really change my filter to use db queries as that would involve a lot of refactoring, any one else have a fix for this?

The patch at https://www.drupal.org/node/2470503#comment-9952629 fixed my issues, hope this helps someone!