please verify

[jelle]

Can someone confirm this on a pristine beta 2 install?
I changed the vertical(bulletmenu) page's parent from menus/navigation systems to  the first entry (home).
This will make breadcrumbs crumble and let apache eat huge resources until it times out. moving the offending page one level up (parent none) makes the page accesible again. moving it back under the first parent breaks again.

Before i spend too much time on digging up this bug, can somebody verify it is indeed a bug?

[Ted]

I'm going to take a guess on this one and say that the hierarchy manager is getting confused for some reason (maybe the timestamp isn't getting reset) and probably goes into an infinite loop.

I'll look into it.  Thanks!

[jelle]

as far as I can see, nothing can be attached to the first menu item.

I never knew programs were couscious yet, so them getting confused confuses me (in soviet russia, to complete the /. style...)

[jelle]

while digging around to see if i can understand this bug, i stumbled on this function in class.contentnode.inc.php. As it always returns -1, what is it's purpose?

Code: Select all

  /**
   * Returns the position of a node into the list of children
   * This method is a workaround for a PHP4 bug where reference testing
   * returns a circular reference fatal error
   * @param $node the node to find into the list of children
   */
  function findChildNodeIndex(&$node) {
    $i=0;
    foreach ($this->children as $child) {
      if ($child->getContent()==$node->getContent()) return $i;
      $i++;
    }
    return -1;
  }

[Ted]

It shouldn't always return -1...

Anyways, I fixed the problem.  It was an issue in breadcrumbs where the set root and the actual root matching would go into a infinite loop.  It's in svn.  thanks!

[jelle]

How often is the svntest code updated? Can that be done automagically?
(jup the last 30 seconds of apache going wild was me testing it there...)

[Ted]

I could probably write a quick cron job to update it regularly.

Anyway, the code is updated.